Developments in Intellectual Property Strategy: The Impact of Artificial Intelligence, Robotics and New Technologies

Edited by
Nadia Naim
Developments
in Intellectual
Property Strategy
The Impact of Artificial
Intelligence, Robotics
and New Technologies
Developments in Intellectual Property
Strategy
“This is a fascinating book that provides its readers with a comprehensive inter-
disciplinary journey to addressing unpacking intellectual property, ethical and
governance challenges while frontier technologies are engaged, as well as explor-
ing responding strategies to the business world. It is an impressive reading book
for academic scholars, business professionals, industry practitioners and
policymakers.”
—Dr Luo Li, Assistant Professor of Law, Coventry Law School, Coventry
University, UK
“Dr. Nadia has put out an awesome literature, which is relevant not just for
academics, but it is for all cadre of society. Her interdisciplinary and cross-
cultural approaches to the treatment of the evolving issues surrounding artificial
intelligence (AI) is uniquely presented in this work: ‘Developments in Intellectual
Property Strategy: The Impact of Artificial Intelligence, Robotics and New
Technologies.’ I highly recommend this book to anyone engaged in the econom-
ics, politics, legal, and technology spaces.”
—Professor Samuel Samiái Andrews, USA Ambassador’s Distinguished Scholar,
Professor of Intellectual Property Law & faculty, member, Prince Mohammad Bin
Fahd University, College of Law, Al Khobar, Kingdom of Saudi Arabia
Nadia Naim
Editor
Developments in
Intellectual Property
Strategy
The Impact of Artificial Intelligence,
Robotics and New Technologies
Editor
Nadia Naim
Law and Social Sciences
Aston University
Birmingham, UK
ISBN 978-3-031-42575-2 ISBN 978-3-031-42576-9 (eBook)

https://doi.org/10.1007/978-3-031-42576-9
© The Editor(s) (if applicable) and The Author(s), under exclusive licence to Springer Nature
Switzerland AG 2024
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Paper in this product is recyclable.

Foreword
The contemporary debates that currently rage in academe and social

media on the nature and merits or otherwise of new technologies, artifi-
cial Intelligence and robotics at times encompass a spectrum ranging
from predictions of a society of positive performance where mankind’s
contribution to climatic chaos is under control, menial labour is per-
formed by machines, poverty and pandemic no longer prevail, to a
robotic-ruled regime and even to the eventual end of mankind as we
know it and think we understand it today.
So sound and reasoned analysis of key elements of the interaction of
intellectual property and its societal and business engagement now and in
any future are at times overlooked in these ranging exchanges.
Dr Nadia Naim endeavours, very successfully in this writer’s humble
view, to introduce much-needed balance in addressing some of these key
elements. Nadia is well qualified to so examine these elements, having
researched and published widely in the field of intellectual property rights
and protection regimes in domestic, regional and broader international
contexts. Her most recent publication, for example, has examined Islamic
legal principles and intellectual property rights in the member states of
the Arabian Gulf states.
v
vi Foreword
This latest book offers chapters by Nadia and a number of expert con-
tributors whose analyses encompass both fundamentals and crucial areas
that are often overlooked or do not receive the deserved attention they
warrant in the raging exchanges over future projections on society and
business. The impact of emerging and future technologies upon society
and its relationship between established current intellectual property
rights protection regimes and the new technologies like artificial intelli-
gence require a sound and balanced contribution to tempering discourses
coloured by speculation and emotion. This book sets out to achieve that
high aspiration. It covers core topics such as artificial intelligence cre-
ations and ownership—who can, and should, own the intellectual prop-
erty that artificial intelligence, as distinct from natural person or legal
entity, might create?—which in itself require examination of the regula-
tion and/or protection of intellectual property creativity or invention
conceived by artificial intelligence or protection and whether such issues
are just a matter of time or a step too far. Irrespective of that response,
society and business are already confronted with the challenge of the big
data and the risk of data legislation such as the EU Data Act and compa-
rable UK statutes enshrining quasi-exclusivity data lockups. The legal
nature of non-fungible tokens, utilising the exemplar of trade in artworks
in the digital environment, also receives attention, as does the potential
for robots to adopt a role as care contributors in the ever-increasing and
concerning industry of proper and appropriate attention to our ageing
population in care. Finally, the book offers perspectives on the value of
intellectual property financing for companies and businesses, and an
examination of transformative (bio)technologies in knowledge societies.
Nadia and her co-contributors bring a fresh and essential critical analy-
sis to the debate on these legal and regulatory considerations of intellec-
tual rights and artificial intelligence within society and the business
world. This book covers a broad yet manageable array of topics, some of
which are otherwise overlooked, on the future of business intellectual
property strategy and planning, with a comprehensive interdisciplinary
Foreword vii
review of the rapidly evolving fields of artificial intelligence and conse-

quent new technologies. It is a welcome contribution and advance to the
current debate.
It will be enjoyed by scholars and practitioners alike, as well as by those
generally of enquiring mind.
Emeritus Professor International Law David Price

School of Law, Faculty of Arts and Society
Charles Darwin University
Darwin, Australia
Preface
Developments in intellectual property strategy is a cutting-edge and

business-focussed publication on the impact of IP in the business world.
Taking an interdisciplinary and diverse perspective, this book enriches
the evolving definition and scope of intellectual property and the impact
of artificial intelligence, robotics and new technologies literature by
focusing on actors, products and regulation that shape the business sec-
tor. Considering the gap between theory and practice, this book bridges
academic and professional knowledge in unpacking legal, ethical and
governance issues in the intellectual property industry. In an effort to
include as many viewpoints as possible, regardless of popularity or who
holds them, the book editor gathered thoughts from diverse fields, includ-
ing business, intellectual property, artificial intelligence, ethics, gover-
nance, law and management. Appealing to academic stakeholders with
an interest in international intellectual property developments and the
impact of emerging technologies, this book is the result of and a testa-
ment to a distinct educational project that includes different countries
and communities for future reference.
The book aims to develop an understanding of the legal challenges
posed by intellectual property, artificial intelligence and robotics tech-
nologies, along with a consideration of appropriate business and regula-
tory responses. It provides a business-focussed framework for considering
concepts and principles that relate to the development and use of such
ix
x Preface
technologies. It considers different legal and regulatory governance

regimes at the international, regional and national levels. It explores the
application and impact of artificial intelligence and intellectual property
in the current Industrial 4.0 revolution era. For artificial intelligence, sce-
narios such as automated and algorithmic decision-making in business
and finance, ageing population and healthcare robots will be discussed.
From the intellectual property angle, the book will examine the pro-
tection of the intangible but marketable proprietary interests related to
goodwill, trademarks, trade and personal secrets, know-how and inven-
tions. In addition it will also examine the protection given to literary,
artistic and musical creations; the products of music recording, film,
broadcasting and other media companies; and designs used in commer-
cial and industrial spheres. This will include the protection given by UK
and EU law and international conventions. The increasing importance of
such law in the modern business environment, the protection of biotech-
nological invention and the impact of the digital environment will be
emphasised. The book will develop a critical awareness of the nature of
the issues surrounding intellectual property rights and the role they play
in the business world. For intellectual property, the chapters will also
cover the value of intellectual property audits for companies, balancing
the need for growth in enterprise and innovation.
The book aims to provide an intellectually stimulating and practically
engaging approach to address legal and ethical issues relating to artificial
intelligence and intellectual property; in particular, Chap. 1 considers
different legal and regulatory governance regimes at the international,
regional and national levels. Currently, all intellectual property rights cre-
ated with human and artificial intelligence “effort” belong to the human
owner; however, as artificial intelligence becomes more sophisticated, the
law on intellectual property protection will need to adapt accordingly.
Chapter 2 considers the legal and ethical questions concerning the
generation and utilisation of big data; the traditional legal distinction
between online-offline issues is not sufficient to deal with the intricacies
of big data, AI and smart technologies and the specialised laws and regu-
lations will be examined on how to regulate issues in specific fields.
Chapter 3 explores the legal nature of non-fungible tokens (NFTs) and
what rights and interests purchasers acquire. It contextualises the
Preface xi
discussion against the background of the trade of artworks in a block-

chain environment; examines the most recent legislation and court deci-
sions concerning NFTs and cryptocurrency; and conducts a comparative
study between the selected common law and civil law systems on their
concept of assets and property in relation to NFTs.
Chapter 4 aims to discuss the involvement of artificial intelligence in
the intellectual property value chain, beyond the earlier questions in the
book on artificial intelligence as a creator/inventor but also the role of
artificial intelligence as a user of protected intangible assets. The chal-
lenges are discussed with respect to two areas of intellectual property:
copyright and patents. There is an in-depth analysis of the DABUS case
and the various judgements in the UK, USA, Germany, South Africa and
Australia.
Expanding further on the theme of artificial intelligence and intellec-
tual property and more specifically on patent law, Chap. 5 discusses the
emerging issues of artificial intelligence in countries beyond those consid-
ered in the earlier chapters. Focussing on Indonesian patent law, which
currently does not include provisions for artificial intelligence inventions,
the analysis highlights the problem in the protection of artificial intelli-
gence inventions, with a deeper study into understanding artificial intel-
ligence inventions and Indonesian patent law. A comparison of studies in
Indonesian and Japanese patent laws on artificial intelligence inventions
is conducted to find similarities and differences between the two coun-
tries regarding the protection of artificial intelligence inventions.
Ageing population and healthcare robots are the focus of Chap. 6.
Scientific progress in robotics, artificial intelligence integrated systems
and increasingly sophisticated software engineering has contributed to
innovative developments in care robots in the Asia Pacific regions, Europe
and the USA. Whilst the use of care robots is not widespread, research is
already occurring to integrate wider use of care robots in the elderly pop-
ulation towards improving their quality of life. Developments in care
robots for the elderly are valuable, yet they give rise to special concerns as
they affect peoples’ health and safety. Key ethical and legal issues arising
from the use of care robots in the ageing population such as deception
and trust from care robots are highlighted, which implicate governance
concerns ranging from balancing innovations and ensuring user safety to
xii Preface
determining the appropriate levels of regulatory oversight and support

for developers and market strategies. An examination of these issues pro-
vides some guidance for law makers and policy developers in these areas
of concern.
An examination of the tangible value of IP financing for companies
and businesses lends weight to the overall vision of the book to weigh in
on the impact of artificial intelligence on intellectual property strategy; as
in a nutshell, the artificial intelligence world wants a slice of the intellec-
tual property pie. Intangible assets are today recognised by many compa-
nies as their most important resource. Without intellectual property
rights, many innovative ventures have nothing to sell or licence. In con-
temporary knowledge-intensive economies, from the world’s largest and
most powerful companies to the smallest small- to medium-sized enter-
prises, the exploitation of intellectual assets—copyright, patents, trade-
marks, designs and know-how—is essential to business and the creative
industries. However intellectual property is rarely sufficient, of itself, to
create businesses or, indeed, to create significant economic value; it needs
to be incorporated into a commercialisation plan. A successful future for
intellectual property financing is a significant step in further develop-
ment of the potential for an intellectual property and artificial intelligence-
based economy.
Chapter 8 aims to gain a better understanding of “knowledge society”
epistemologies and explore the role of patents in cellular agriculture, a
field of enquiry that uses cell-cultivation technology, as a case study to
elucidate the extent to which intellectual property rights can be deployed
to generate optimal public welfare. Through a public interest lens, it also
seeks to understand whether growing calls for open science can be aligned
with the needs of a flourishing innovation ecosystem. Combining theo-
retical and doctrinal legal approaches along with insights from political
and economic theories on regulation and governance, the final chapter
looks at some of the legal questions to illuminate how governments, regu-
lators and stakeholders balance and meet the demands of pressing social
challenges, such as climate change and food insecurity, with the benefits
of transformative biotechnologies to create intellectual commons.
Birmingham, UK Nadia Naim

Contents

Artificial Intelligence Creations and Ownership – Who Should
the Intellectual Property Belong To? 1
Nadia Naim
1 What Is Artificial Intelligence (AI)? 3
2 Defining Artificial Intelligence Within Intellectual
Property Law 4
3 EU Commission and Artificial Intelligence 6
4 So Where Does That Leave Us? 7
5 To DABUS or Not to DABUS 8
6 The Copyright and AI Battle 9
7 Can the Four-Step Test Be Applied to AI-Generated Outputs
and Pave the Way for AI Ownership of Copyright? 12
8 Non-human Actors in the Originality Test 14
9 Artificial Intelligence: A Black Swan for Intellectual Property
Systems? 15
10 Legal and Ethical Considerations 17
11 Trustworthy AI 17
12 Patent Offices and AI 18
13 AI in Copyright Law 19
14 AI in Patent Law 21
15 Alas, Who Should the Intellectual Property Belong To? 21
References 23
xiii
xiv Contents

Fostering Innovation by Utilising Big Data: The Data Act and
the Risk of Quasi-Exclusivity Reinforcing Data Lockups 25
Marc A. Stuhldreier
1 Introduction 25
2 Utilising the Value of Big Data: Innovation in the Fourth
Industrial Revolution (4IR) 27
2.1 Big Data 27
2.2 The Importance of Big Data for Future Innovation in
the 4IR 30
2.3 Unlocking the Value/Potential of Big Data 31
3 The EU Data Strategy and the Data Act 38
3.1 The European Data Strategy 38
3.2 The Data Act 40
4 A de-Facto Exclusive Right to Data 53
4.1 The Problem with Exclusivity: Lessons from the
Global Harmonisation of Intellectual Property Rights 55
4.2 An Exclusive Right to Data and the Problem of IP
Overlaps 57
5 Conclusion 58
References 60

Legal Nature of NFTed Artwork: A Comparative Study 65
Jia Wang and Arianna Alpini
1 Introduction 65
2 Blockchain and the Tokenisation of Artwork 68
2.1 The Blockchains: Private, Public and Consortium Chain 68
2.2 The Tokens: Fungible and Non-fungible 69
3 The Trading of Artworks 70
4 Legal Status of NFT: The Common Law Approaches 73
4.1 The United Kingdom 73
4.2 Singapore 75
5 Legal Status of NFT: The Civil Law Approaches 76
5.1 The Theoretical and Legislative Discussion 76
5.2 European Cases 78
5.3 Chinese Cases 81
6 Conclusions 82
References 85
Contents xv

Intellectual Property Regulation of Artificial Intelligence: A
Matter of Time or a Step Too Far? 91
Lucius Klobucnik
1 Introduction 91
2 AI and Copyright 93
2.1 Originality of Works Created by AI 93
2.2 Potential Legislative Lead Given by Computer-
Generated Works 95
3 An Appropriate Mechanism to Use Copyright-Protected
Works by AI 96
3.1 AI Creator Learning from Existing Copyright-
Protected Works 96
3.2 Text and Data Mining (TDM) Exception 97
4 DABUS Case – A Milestone in Patent Case Law or an
Unsuccessful AI Challenge to Patent Law 101
4.1 AI and Patent Law 101
4.2 United Kingdom 101
4.3 USA 104
4.4 Germany 104
4.5 South Africa 106
4.6 Australia 106
5 Conclusion 107
References108
Artificial Intelligence Invention Protection Model113

An
Budi Agus Riswandi
1 Introduction 113
2 Discussion 118
2.1 Legal Protection of AI Invention Based on Indonesia
Patent Law 118
2.2 Comparative Studies on Protection of Artificial
Intelligences Invention Between Indonesia, United
States, and Japan Patent Law 120
2.3 An Artificial Intelligence Invention Protection Model 121
xvi Contents
3 Conclusion & Suggestion 125

3.1 Conclusion 125
3.2 Suggestion 126
References127

Care Robots for the Elderly: Legal, Ethical Considerations
and Regulatory Strategies129
Hui Yun Chan and Anantharaman Muralidharan
1 Introduction 129
1.1 An Ageing Population 129
1.2 Ethical Complexities 133
1.3 Deception and Manipulation 134
1.4 The ‘Zombie Relationship’ Problem 137
2 Governance Framework for Care Robots 141
2.1 Legal Complexities and Considerations 141
2.2 User Health, Safety and Wellbeing 143
2.3 Data Security and Privacy Protection 143
2.4 Balancing Innovations, Market and Protecting User
Safety145
2.5 Regulatory Strategies and Options 148
3 Conclusion 151
References152
Examination of the Tangible Value of IP Financing for

An
Companies and Businesses157
Nadia Naim
1 IP and Business 158
1.1 What Is an IP Audit? 159
1.2 Processes to Identify, Capture and Manage IP 160
1.3 Contents of an IP Audit 161
1.4 Patents Audit 162
1.5 Design Rights Audit 162
1.6 Trademarks Audit 163
2 Passing Off 165
3 Measuring the Value of Goodwill 166
3.1 Hidden Assets 167
3.2 Copyright Audit 167
Contents xvii
4 IP Due Diligence 168

4.1 How Does IP Generate Revenue? 169
5 IP Valuation Approach 169
5.1 The Cost Method 171
5.2 The Income or Economic Benefit Method 172
5.3 The Market Value Approach 173
5.4 Exploiting Your IP 174
5.5 IP Financing and Emerging Technologies 176
6 In Conclusion: IP Financing for Businesses 177
References178

Transformative (Bio)technologies in Knowledge Societies:
Of Patents and Intellectual Commons181
Mariela de Amstalden and Nivita Sukhadia
1 Transformative Biotechnologies: Cell-Cultivation for
Human Food Consumption 181
2 Saving the Intellectual Commons in Future Knowledge
Societies183
3 Patent Effectiveness and the WTO TRIPS 184
4 Conclusion 188
References188
A
nnexure: Table of Cases191
I ndex193
Notes on Contributors
Abdurrahman Alfaqiih is a lecturer in the Department of Civil Law also the

director of the Intellectual Property Rights Centre at the Faculty of Law,
Universitas Islam Indonesia, Indonesia. He holds a PhD in Law at the University
of Groningen, the Netherlands. He is experienced in journal publications in
national and international journals; he has also written a book with his team.
His publications, books and research works are related to intellectual property
rights law and technology.
Arianna Alpini is Associate Professor of Private Law in the Department of Law
at the University of Macerata, Italy, where she lectures on private law and sport
law. She also teaches foundations of private law in the European and Comparative
Legal Studies international degree programme. She was Assistant Professor of
Private Law and Adjunct Professor of Legal Method at University of Sannio,
Italy. She is a fellow of the European Law Institute (ELI) and a member of the
Italian Association of Civil Law Scholars (SISDiC). Her research interests
include the impact of technologies on contract law. She is an author of mono-
graphs, essays, articles and notes on obligations, co-ownership, European prin-
ciples, interpretation and sources of law and fundamental rights protection. She
is the principal investigator in a research project on transdisciplinary methodol-
ogy, and the coordinator of a research unit on consumer protection.
Muralidharan Anantharaman graduated from National University of
Singapore (NUS) with a BSc in Life Sciences in 2010 and then an MA in
Philosophy in 2014. In 2014 and 2015 he was a teaching assistant in the
xix
xx Notes on Contributors
Department of Philosophy and a research assistant at the Centre for Biomedical

Ethics; he holds a PhD in Philosophy from the University of Warwick, UK. After
graduating in 2020, he became a lecturer at Singapore University of Social
Sciences (SUSS), before re-joining the Centre as a research fellow in 2021. He
has research interests in political philosophy, epistemology as well as normative
and applied ethics.
Hui Yun Chan is a Research Fellow in the Yong Loo Lin School of Medicine at
the Centre for Biomedical Ethics, National University of Singapore, Singapore.
Hui Yun has held several law lectureships in universities in the UK and is a fel-
low of the Higher Education Academy, UK. She has written scholarly articles in
the fields of bioethics and health law on topics in healthcare decision-making
law and ethics, organ donation, and a monograph on advance healthcare
decision-making titled Advance Directives: Rethinking Regulation, Autonomy and
Healthcare Decision-Making (2018). She has written about the legal and ethical
aspects of COVID-19 related issues, and public health regulation and ethics.
Mariela de Amstalden’s research explores the legal implications of emerging
biotechnologies, in particular animal cell cultivation, and their implications for
international economic law and global governance. Her monograph, Global
Food Governance: Implications of Food Safety and Quality Standards in International
Trade Law (Berne, 2015), explores how the notion of legal standards in interna-
tional law can be interpreted to ensure that global food demands are met. She
has experience in private practice and the judiciary and has written a number of
peer-reviewed articles about the law of emerging biotechnologies, food and pub-
lic health regulation at the intersection with international economic law and
global governance.
Dodik Setiawan Nur Heriyanto is a senior lecturer in the Department of
International Law at the Faculty of Law, Islamic University of Indonesia,
Indonesia, since 2012. He is the founder and chair of Base for International Law
and ASEAN Legal Studies (BILALS) in Indonesia. He is very active in conduct-
ing research, with particular interests in humanitarian law, diplomatic and con-
sular law and international dispute settlement law.
Lucius Klobucnik is Lecturer in Intellectual Property Law at Aston University,
Birmingham, UK. He holds a PhD degree from Queen Mary University,
London, and Max Planck Institute for Innovation and Competition in Munich
(Augsburg University). His main research focuses on copyright law (especially
music rights), intellectual property in new technologies and platform liability.
Notes on Contributors xxi
As part of his PhD research project, he worked at the German Collective

Management Organisation in Music Rights (GEMA) in Munich, and at CISAC
(Confederation of Societies of Authors and Composers) in Paris.
Nadia Naim is Associate Dean for Internationalisation and Senior Lecturer in
Law at Aston University, UK. She is a qualified barrister and fellow of the Higher
Education Academy. She has written many scholarly articles in the areas of inter-
national intellectual property law, intangible asset financing and development,
comparative law, Islamic law and international law. She has authored books in
the intellectual property field and specialises in emerging areas of intellectual
property and value-based financing development.
Budi Agus Riswandi is Professor of Law in the field of intellectual property
rights and cyberlaw and a dean of the Faculty of Law, Universitas Islam Indonesia,
Indonesia. He teaches intellectual property law at the Faculty of Law, Universitas
Islam Indonesia. In addition, he also teaches at various law faculties in Indonesia.
He usually conducts research focusing on the field of intellectual property rights
and cyberlaw and publications in the form of books and journals in the field of
intellectual property rights and cyberlaw. He is also usually appointed as a
speaker in various scientific activities such as seminars or conferences both at
national and at international levels, especially in the field of intellectual property
rights and cyberlaw. He is also registered as an advocate with the Indonesian
Advocates Association and a registered Intellectual Property Consultant at the
Directorate General of Intellectual Property of the Ministry of Law and Human
Rights of the Republic of Indonesia as well as the Coordinator of the Intellectual
Property Consultants Association of Yogyakarta Special Region. In addition, he
is also the Chairman of the Indonesian Intellectual Property Centre Association.
Marc A. Stuhldreier is a Postdoctoral Fellow at Linköping University, Sweden,
in the ERC-funded PASSIM project. He is a qualified data protection officer
and auditor in Germany. His research focusses on the intersection between intel-
lectual property rights and human rights, with a particular focus on the impacts
of private exclusive rights on public access. Focussing on the accessibility of
medicines in much of his works, he further addresses new related issues stem-
ming from currently debated data regulations and their anticipated impact on
future public research and innovation.
Nivita Sukhadia holds a LLB in Law with Business Studies graduate from the
University of Birmingham, UK, and is also a Master of Law student at the
University of Cambridge. She works as a part-time research assistant in the intel-
lectual property law field.
xxii Notes on Contributors
Angelia Jia Wang is an assistant professor in the Law School at Durham

University, UK. Prior to joining Durham Law School in 2020, she taught at the
Hong Kong Polytechnic University. She has been a Research Fellow at the
Berkman Center for Internet and Society (now the Berkman Klein Centre),
Harvard University, and a Postdoc Fellow at the Law School, Singapore
Management University. Her research interests lie in the areas of intellectual
property law and the intersection between law and technology. Her recent
research projects include intellectual property law (IP) and artificial intelligence,
IP and video games and IP for tokenised artwork. Her work is published in
European Intellectual Property Review, Intellectual Property Law & Practice, Hong
Kong Law Journal, European Review of Private Law and Asian Pacific Law Review,
and she has also written a monograph.
Artificial Intelligence Creations
and Ownership – Who Should
the Intellectual Property Belong To?
Nadia Naim
Research in this area enters the rapidly growing artificial intelligence and
robotics industries in the legal, business, manufacturing, and healthcare
sectors and the impact of intellectual property protection on emerging
technologies. This chapter aims to develop an understanding of the legal
and ethical challenges posed by artificial intelligence and robotics tech-
nologies, along with consideration of appropriate legal and regulatory
responses. It provides a philosophical and legal framework for consider-
ing concepts and principles that relate to the development and use of
such technologies. It considers different legal and regulatory governance
regimes at the international, regional, and national levels. Currently, all
intellectual property rights created with human and artificial intelligence
“effort”, belong to the human however as artificial intelligence becomes
more sophisticated, the law on intellectual property protection will need
to adapt accordingly. The chapter will focus on the interplay between
N. Naim (*)
Law and Social Sciences, Aston University, Birmingham, UK
e-mail: n.naim@aston.ac.uk
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 1

N. Naim (ed.), Developments in Intellectual Property Strategy,
https://doi.org/10.1007/978-3-031-42576-9_1
2 N. Naim
intellectual property and artificial intelligence, intellectual property rights

protection affords the human intellectual property rights holder a time-
limited monopoly over their intangible asset and has yet to afford any
mirror rights or alternative rights to artificial intelligence.
At present, artificial intelligence can be seen to adequately respond to
humans’ needs, such as virtual digital assistance, most people will have
Apple Siri, Alexa and or Google Home, that respond using voice recogni-
tion systems. Many rely on artificial intelligence to provide a recom-
mended list of music based on your existing music choices and preferences,
and produce works of art (Li 2020). Most notably, a project team behind
The Next Rembrandt designed algorithms that allowed a computer to
create a painting in the style of the seventeenth century Dutch artist and
is known as the Rembrandt 2.0. Artificial intelligence can produce works
which could be considered as copyright works however the law has yet to
acknowledge AI as a copyright owner. Humans working in creative, inno-
vative and legal sectors are discussing the consequence of AI systems
when it comes to who will own the intellectual property, more impor-
tantly, who will the economic rights belong to. Artificial intelligence sys-
tems are developing at a significant pace and as a result, reshaping the
whole creative and innovative sectors that are protected in the existing
intellectual property system. Therefore, it is necessary to identify the AI
systems at present, defining and distinguishing between the concepts of
“AI-assisted” and “AI-generated”, to outline the direction of AI develop-
ment in the context of intellectual property law.
Although the advantages of AI in our daily lives are undeniable, there
are concerns about its dangers. Inadequate physical security, economic
losses, and ethical issues are just a few examples of the damage AI could
cause. The European Commission has identified examples of unaccept-
able risk, such as uses of AI that manipulate human behaviour and sys-
tems that allow social-credit scoring. For example, the European legal
framework would prohibit an AI system that resembles China’s social
credit scoring. From an intellectual property perspective on AI, the World
Intellectual Property Office (WIPO) leads several conversations on
Intellectual Property and Frontier Technologies, with the European
Union working on a legal framework to regulate artificial intelligence and
Artificial Intelligence Creations and Ownership – Who Should… 3
the UK Intellectual Property Office (UK IPO) leading several consulta-

tions on ownership.
1 What Is Artificial Intelligence (AI)?

Marvin Minsky defined AI as intelligent behaviour, saying “AI is the sci-
ence of making machines do things that would require intelligence if
done by men” (Minsky 1961). He omitted to mention women but as the
founding figure in AI, we can forgive the omission. This is now consid-
ered weak AI as with the speed of developments, there is a demand for
stronger AI that not only requires intelligence but identically mirrors
human intelligence. David Fogel (2005) went further to identify strong
AI as “the study of intelligent behaviour. Its ultimate goal is a theory of
intelligence that accounts for the behaviour of naturally occurring intel-
ligent entities and that guides the creation of artificial entities capable of
intelligent behaviour”.
The three waves of AI were described in a presentation by DARPA’s
John Launchbury (2017). He set out how artificial intelligence capabili-
ties can be divided into three distinct waves of; handcrafted knowledge
(describe), statistical learning (categorise) and contextual adaption
(explain). Launchbury’s definition of first-wave AI systems are rule-based
systems and human engineers define the rules for AI systems to follow.
An example is Turbo Tax that tax lawyers or accountants use to convert
the complex tax codes into rules, these AI systems do not have learning
capability, are poor in handling any uncertainties and can only solve nar-
rowly defined problems. Therefore, it can be said the first wave of AI
systems are simply machines following rules that are defined by humans
with no substantial difference from electronic tools or equipment that
humans use. The second-wave of AI systems have capabilities to define
rules through clustering and classifications of massive data – this is sig-
nificantly different from the first-wave of AI systems and also distinguish-
able from the third-wave of contextual adaption.
A more recent definition comes from Waymond Rodgers (2020) where
he describes three waves of AI, namely narrow, general and super, which
builds on the definitions by Launchbury. Narrow AI has only
4 N. Naim
characteristics consistent with cognitive intelligence, it is not conscious,

sentient, or driven by emotions the way that individuals are configured to
make decisions. General AI represent machines that display human intel-
ligence, that is, they can perform any intellectual task that a human being
can in terms of decision-making. Super AI displays features of all kinds of
competencies such as emotional and social intelligence, can be self-
conscious and self-aware in their interactions with others.
2 Defining Artificial Intelligence Within

Intellectual Property Law
AI systems and their advanced development are reshaping the whole cre-
ative and innovative sectors that are protected in the existing intellectual
property system. Therefore, it is necessary to identify the AI systems at
present and to map the direction of AI development in the context of
intellectual property law. World Intellectual Property Organisation
(WIPO) leads a series of conversations and discussions in AI and intel-
lectual property policy and through the Conversations, WIPO tries to
clarify the line between the two concepts: AI- assisted and AI-generated
outputs, because they would lead to substantially different recognitions.
The division of AI outputs for the purposes of AI can be better seen as
three tiers and defined as “AI- supported”, “AI-assisted” and “AI-generated”.
The least contentious of the three tiers is “AI-supported”. This is since
the results made by AI-supported systems are predictable by humans.
From this perspective, AI-supported outputs are defined as outputs pro-
duced by a rule-based computer system that fully follows the rules defined
by humans, without the substantial intelligent ability (which embraces
learning, perceiving, abstracting and reasoning). In this case, humans are
authors/inventors of AI-supported outputs because the outputs are a
result of humans’ full material intervention and therefore owners of the
intellectual property. The AI-support in the creative process is limited to
a digital tool for scientists, entrepreneurs and artists, enabling new human
inventions and creations (UK IPO 2022).
The second-wave of AI systems, which can be clearly seen as AI- assisted

over AI- supported, relies on the advanced improvement of data and
computational power as well as better models of algorithms than seen at
the AI-supported wave or handcrafted knowledge. The rising big data
and internet technology builds up a giant database, which is a perfect
machine learning hub. The more data the AI systems can obtain, the
wider knowledge AI systems can learn from and therefore the capability
to produce more accurate outputs is enhanced. The second pillar of the
success of AI systems is computational power, which means computa-
tional capacity. Meanwhile, the present AI systems can use those defined
models to predict and make decisions by themselves to some extent.
We can reflect on the first two waves of AI, the first of which was hand-
crafted knowledge, which still has relevance. The second wave, which is
very mainstream with AI systems for face and voice recognition, is
focussed on statistical learning where we build systems that get trained on
data. However, the two waves by themselves are not going to be sufficient
or meet the growing market demand for AI systems and what they can
do. We see the need to bring them together and that is evident through a
third wave of AI technology built around the concept of contextual adap-
tion (Rodgers 2020).
The term “AI-generated outputs” is defined as outputs made by com-
puter systems that have the freedom to do decision-making during the
output generation process through integrating the ability to perceive,
abstract, explain and reason by themselves, with minimised (if any)
human intervention and embracing their own understanding of the real
world. Here, we have the contention between intellectual property pro-
tection and AI acknowledgment in the intangible asset industry. There
are strong arguments that arts and music are inherently human and as
such the associated intellectual property rights can only belong to the
human creator. There are several issues with this, first of all, there is little
incentive for the AI industry to further develop the third wave of the AI
systems if it cannot benefit from intellectual property rights protection.
To explain this further, the EU Commission stance on the extension of
intellectual property rights to AI and the relationship between each indi-
vidual intellectual property right and AI will be considered.
6 N. Naim
3 EU Commission and Artificial Intelligence

The European Parliament resolution of 16 February 2017 made recom-
mendations to the Commission on Civil Law Rules on Robotics (EU
Commission on Civil Law Rules on Robotics 2015), followed by a reso-
lution of 20 October 2020 on intellectual property rights for the devel-
opment of artificial intelligence technologies. The EU Commission has
proposed the first ever legal framework on AI, which addresses the risks
of AI and positions Europe to play a leading role globally (European
Commission, 2021). MEPs approved proposals to address long-term
opportunities and legal challenges posed by AI in the area of ethics, civil
liability and intellectual property. The question is what is an effective
intellectual property system that safeguards innovation and creation
where AI applications are involved? There are currently no uniform con-
tractual clauses used by all or most AI service providers.
The Commission report states that EU global leadership in AI requires
an effective intellectual property system and safeguards for the EU’s pat-
ent system to protect developers. MEPs specify that AI should not have
legal personality, and “inventorship” should be only granted to humans.
a test case before the UK courts (as well as before courts and IP offices
globally), in the UK – Thaler v Comptroller General of Patents Trade
Marks And Designs [2021] EWCA Civ 1374, has confirmed the position
under the current law. All judges in the Court of Appeal agreed that an
AI machine cannot be an inventor under the 1977 Patents Act. There was
a disagreement, however, as to whether a patent application which con-
tained a statement of inventorship to that effect was valid under the Act,
with the majority concluding that it was not.
Members stressed that creating a framework for creativity and innova-
tion by encouraging the use of AI technologies by creators should not be
at the expense of the interests of human creators or the Unions ethical
principles. They considered it essential in this respect to distinguish
between AI-assisted human creations and AI-generated creations. They
specified that AI should not be endowed with legal personality, which
could have negative effects on the motivation of human creators. Members
therefore recommended that rights should only be granted to natural or
legal persons who have created the work legally and only if the copyright
owner has given permission for the use of copyrighted content. The reso-
lution added that AI or related technologies used for the registration pro-
cedure to grant IPRs and for the determination of liability for
infringements of IPRs cannot be a substitute for human review carried
out on a case-by-case basis, in order to ensure the quality and fairness of
decisions.
4 So Where Does That Leave Us?

Let’s start with patents. WIPO set the minimum compliance standards
for all intellectual property rights that all signatory countries must abide
by, knows as trade related intellectual property rights (TRIPS). For pat-
ent registration and subsequent protection, the rules set out a four-
step test:
(1) The invention is new

(2) An invention shall be taken to involve an inventive step if it is not
obvious to a person skilled in the art
(3) An invention shall be taken to be capable of industrial application if
it can be made in any kind of industry
(4) It does not fall foul of the exclusions and exceptions
All WIPO signatories by in large follow the same four step test, indi-
vidual countries will differ on what they deem as an exclusion or excep-
tion dependent on the Constitution and commonly these fall under
morality or public policy. A patent is an official document that confers
proprietorship of an invention on the recipient, grant of a patent is pre-
ceded by examination of applications by the patenting authority. The
fundamental principle behind patents is that the government awards
exclusive control over an invention for a fixed number of years to the
individual who first discloses the invention within its territory.
In most systems, a patent is granted to an applicant who is first to
submit a detailed description of the invention, for example in the UK,
the system operates on a first to file model. The TRIPS agreement sought
to set international minimum standards in patent protection and at
8 N. Naim
Article 27 stated the requirements for; novelty, inventiveness, industrial

applicability and exclusions. A successful patent application grants the
inventor the right to exclude others from utilising the patented invention
for a given amount of time, under the TRIPS Agreement, the monopoly
right is granted for twenty years. In return, the inventor is under an obli-
gation to describe the invention in detail to give notice to the public. The
counterargument to the monopoly right granted to the inventor is that
monopoly is against public interest.
5 To DABUS or Not to DABUS

In oral proceedings in December 2021, the Legal Board of Appeal of The
European Patent Office (EPO) presided over the case of DABUS and
dismissed the appeal in cases J 8/20 and J 9/20.
The Legal Board of Appeal confirmed the decisions of the Receiving
Section of the European Patent Office to refuse the applications EP 18
275 163 and EP 18 275 174, to not recognize DABUS, an artificial intel-
ligence system, as the inventor on the patent application. The EPO there-
fore came to the decision that AI cannot be named as the inventor on
patent applications and confirmed previous decisions whereby an inven-
tor in a patent application must be a human being under the European
Patent Convention (EPC).
The dismissal of the DABUS case by the EPO Legal Board of Appeal is
just one of the appeals brought forward by Stephen Thaler, to give intel-
lectual property ownership rights to an AI system as the inventor on a pat-
ent application. This is not Mr. Thaler’s only court battle to have DABUS
recognized as a patent inventor. He has had simultaneous applications in
several countries. For example, in the UK, in September 2021, the Court
of Appeal also dismissed the appeal and upheld that the protection afforded
to human inventors cannot be extended to AI systems. However, although
the UK and EPO are effectively singing off the same IP hymn sheet, Mr.
Thaler has been successful in South Africa (Cochrane and Mhangwane
2022). Therefore, the more pressing question raised by the DABUS case is
whether the intellectual property laws that have been designed to protect
creative human endeavors, can be extended to AI systems.
For patent protection, on the whole, as it currently stands the answer

is no. This links back to the focus of this chapter, who should own the
intellectual property asset and reap the economic benefits generated by
the intellectual effort. There are inherent weaknesses in the intellectual
property system and TRIPS in general as there is no clear provision for
what an AI system needs to demonstrate to meet the requirements of pat-
ent protection. When the distinction between human intellectual effort
and AI effort becomes more difficult to define, there will be repeated calls
from a whole generation of Mr. (and Mrs) Thalers’. As AI develops
through the different cycles of AI assisted and AI generated, the next
cycle will be far more sophisticated. It is important to note, that the sev-
eral consultations carried out by IP offices on the ownership of patent
rights by AI systems don’t recommend a like for like patent ownership
test. There are obvious reasons for this, the speed at which advanced AI
systems could meet the current patent test could be much faster than the
human inventor and consequently, the duration times for any protection
that may be afforded in the future would need to reflect this.
6 The Copyright and AI Battle

Looking at copyright on the other hand, it is evident that the rules are less
clear when compared to patents and AI applications, AI is heavily used in
the music industry and could impact copyright laws. Questions that need
to be considered are what does AI have to offer music producers, and is it
a boon or a threat to people who make music for a living? What does it
mean for copyright protection?
Turning attention to the UK copyright laws, the Copyright Design
Patent Act (CDPA 1988) at section 9(1) defines the ‘“author”, in relation
to a work, as the person who creates it, therefore a natural person.
Copyright is automatic as established by the idea expression dichotomy
from TRIPS article 9(2). The test for copyright is that it must be original
and this is a fairly low benchmark. Most importantly, the copyright work
must fit into a category and some level of effort, a sufficient level of skill,
labour and judgement is required as well as the requirement that it can
be fixed.
10 N. Naim
After several consultations on Copyright and IP, the UK IPO has con-
firmed that a copyright work can be created by a human who has assis-
tance from AI. If the work expresses original human creativity it will
benefit from copyright protection like a work created using any other
tool. By default then, if the work is AI assisted or generated and lacks
original human creativity, it cannot be classed as a copyright work.
Exclusion apply to computer generated works as a sui generis right.
CDPA section 178 a computer generated work with no human author,
the “author” of a “computer- generated work” (CGW) is defined as “the
person by whom the arrangements necessary for the creation of the work
are undertaken”. The protection lasts for 50 years from the date the work
is made.
Distinction between when there is human involvement to when there
isn’t needs to be made more clear which can be achieved with a tiered
distinction between AI assisted and AI generated in all categories of copy-
right, not only computer generated works.
The extent and nature of human involvement needs to be specified.
Does it suffice if a human instructs an AI application to produce music
of a specific genre? Is that sufficient to make the resulting work fall out-
side the qualification of AI generated work? This is relevant to authorship
and ownership of the work. Thinking about the creative process in
AI-assisted outputs, who does the creative process in AI-assisted started
with, what is the equivalent production for the idea expression dichot-
omy (Fig. 1).
Fig. 1 The idea expression dichotomy

There are several stages identified when assessing AI assistance in copy-

right works starting with the conception stage. Under AI assisted out-
puts, this would be the human creator as the designer, working on the
specifications, elaborating a work at this stage. Next, there is an execution
stage where draft versions are produced and this is where the work is pre-
dominantly carried out by the AI system. The final stage of the creative
process reverts back to the human creator as redaction require editing and
finalising the work. For example, a musician using an AI music composer
will often edit the output before releasing the music as a final work. While
AI systems play a dominant role at the execution phase, the role of human
authors at the conception stage often remains essential.
Moreover, in many instances, human beings will also oversee the redac-
tion stage. Depending on the facts of the case, this will allow human
beings sufficient creative choice. Assuming these choices are expressed in
the final AI-assisted output, the output will then qualify as a copyright-
protected work. By contrast, if an AI system is programmed to automati-
cally execute content without the output being conceived or redacted by
a person exercising creative choices, there will be no work, it will be con-
sidered AI generated and in the public domain.
The EU Commission report identified a four-step test for the assess-
ment of copyright protection of subject matter as a “work” where AI
plays a part. (from the EU Commission IP Action Plan).
The test is:
(1) a “production in the literary, scientific or artistic domain”;

(2) the product of human intellectual effort;
(3) originality – the result of creative choices; and
(4) expression – the creative choices “expressed” in the output.
Artificial intelligence in the music industry is a topic of interest, given

the level of dependency human music creators place on artificial intelli-
gence systems to support the creation of the musical work. The economic
value of music copyright globally is estimated at $40 billion (Cooke
2022) and the role artificial intelligence plays in the creation of copy-
rightable work, leads to legal and policy implications in this field.
12 N. Naim
The Institute for Information Law (IViR) and Centre for International
Intellectual Property Studies (CEIPI) carried out a survey looking at
music and AI outputs and found many of the AI music outputs examined
will likely pass steps (1), (2) and (4). The crux of the test is therefore in
step (3), the core of the originality standard under copyright law. From
this perspective then, where an output does not qualify as original in the
sense that it reflects the author’s free and creative choices, that output is,
from the perspective of copyright, in the public domain. Although it
could still benefit from protection under related rights, where is the
incentive for artificial system creators to further enhance and improve the
offering of systems if the current copyright laws exclude the AI generated
work from protection (Bulayenko 2017).
In the context of step (3), it is possible in the first place to identify a
series of external constraints on the assessment of originality: rule- based,
technical, functional, and informational. The existence of such con-
straints reduces the author’s margin for creative freedom, sometimes
below the originality threshold. In the second place, the step allows for
the identification of three iterative stages of the creative process when
using an AI system: “conception”, “execution”, and “redaction” as dis-
cussed above. The iterative stages could support the qualification of AI
outputs for copyright protected work. (Grauwe and Gryspeerdt 2022).
7 Can the Four-Step Test Be Applied

to AI-Generated Outputs and Pave
the Way for AI Ownership of Copyright?
In the landmark Infopaq case (Infopaq International 2009) heard before
the Court of Justice of the European Union (C- 5/08 Infopaq International
A/S v Danske Dagbaldes Forening), that Court held that copyright only
applies to original works that reflect the author’s own intellectual creation
as seen in the copyright four step test. The decision upheld the conserva-
tive view that a human author is necessary for a copyright work to dem-
onstrate and meet the necessary intellectual creativity requirement.
Despite the requirement for human intellectual creativity, the chal-

lenges to copyright laws where artificial intelligence is concerned contin-
ues. In 2016, The Rembrandt 2.0 was created as a computer-generated
artwork. The pertinent question raised by the Rembrandt 2.0 is on who
should own the copyright and is a useful benchmark for the four step AI
copyright test. Let’s first look at the work itself and how it was generated.
The aim of work was not to create a specific painting per se, but rather, to
utilise the Artificial Intelligence systems at the disposal of the creative
team and create a simulation. An extensive range of expertise were gath-
ered in the creation of the Rembrandt 2.0. ING and the J. Walter
Thompson agency in Amsterdam commissioned the project with the use
of a computer from Cornell University researchers that had taught itself
physics and the Deep Dream Generator from Google for surreal photo-
graphs. Algorithms were designed with traditional data analysis systems
to create a digital artwork painting that followed the artistic style and
creativity of the seventeenth century Dutch artist (Guadamuz 2017).
The goal of ING and the J. Walter Thompson agency was to discover
if an algorithm could be created to produce a physical work of art that in
essence, would be the nearest effort to a genuine Rembrandt painting.
The Rembrandt team, consisting of data scientists, engineers, and art his-
torians analysed over 300 Rembrandt’s painting and his techniques, style
and subject matter. 150 gigabytes of digitally rendered graphics were col-
lected to provide the instruction set to produce the textures and layers
necessary for Rembrandt 2.0 and transfer that knowledge into the soft-
ware which could generate the new work using the latest in 3D printing
technology. Who then should the copyright belong to and are the current
copyright laws and frameworks sufficient?
Applying the four-step test to Rembrandt 2.0, for step 1, the work is a
digital painting and therefore fits the category of a production in the
artistic domain. For step 2, the artwork has to demonstrate that it is a
product of human intellectual effort which would be the Rembrandt
team of experts and their effort and step 4 is met by the expression of the
output. Focussing then on step 3, we know AI can support the creative
process provided the originality test is met by a human, copyright law has
yet to develop a test that distinguishes between when the AI is a tool in
the creative process or actually the decision maker that meets the origi-
nality test.
14 N. Naim
J. Walter Thompson Amsterdam (2016)
8 Non-human Actors in the Originality Test

To have AI as a non-human actor in the originality test has both advan-
tages and disadvantages. The clear disadvantage is that the rule is too
broad brushed, essentially placing AI systems and monkeys in the same
category. In the Compendium of US Copyright Practices § 313.2, the
section clearly states what cannot be registered for copyright protection
and includes a photograph taken by a monkey as well as a machine. For
one, the monkey selfie was a rare and complete stroke of luck photo-
graph, the monkey could not satisfy the test for copyright protection and
therefore the US Court correctly held that the image could not belong to
the monkey.
On the US Copyright Practices on machines, the section states:
The Office will not register works produced by a machine or mere mechan-
ical process that operates randomly or automatically without any creative
input or intervention from a human author. The crucial question is

“whether the ‘work’ is basically one of human authorship, with the com-
puter [or other device] merely being an assisting instrument, or whether
the traditional elements of authorship in the work (literary, artistic, or
musical expression or elements of selection, arrangement, etc.) were actu-
ally conceived and executed not by man but by a machine. (U.S. Copyright
Office 1966).
This links back to the initial argument of this chapter as to whether the
current divide between AI assisted and AI generated requires further clar-
ification or better yet, whether a third wave of AI will address step 3 of
the copyright test and be able to fulfil all three iterative stages of the origi-
nality test.
9 Artificial Intelligence: A Black Swan

for Intellectual Property Systems?
Intellectual property (IP) law was treated as a law stimulating creation
and innovation through rewarding human creative and innovative activi-
ties so as to encourage continuing investment in creation and innovation.
Artificial intelligence (AI), being a frontier technology, is challenging the
existing IP framework and testing boundaries of fundamentals of IP sys-
tems. First of all, AI engages creativity and innovation that were treated
as human-only areas. Therefore, there is a debate on whether AI and its
outputs should be protected, can be protected, and how to treat these
issues via existing IP context. Furthermore, the concept of stimulating
creation and innovation through a rewarding system is not suitable for AI
as machines would actively produce creative/innovative outputs without
any rewarding purpose. In this case, it is worthy to understand whether
the existing IP system would stimulate AI- related innovation and invest-
ment and what kind of legal system is able to offer such effort in a wider
legal context. The focus now will be to analyse the AI challenges that IP
systems face and explore how the law might respond to these challenges
in the purpose of stimulating AI investment and development. For there
to be another wave of AI, that is beyond a contextual adaption of the
16 N. Naim
AI-assisted and AI-generated waves, there needs to be an IP system in

place that recognises the AI input, otherwise apart from the contractual
benefits for producing more sophisticated AI, there is little reward from
an IP perspective.
There is considerable economic investment that goes into the develop-
ment of AI systems and if there are no IP laws to protect the work that is
created, it must then be free to use in the public domain. Therefore,
despite AI being the black swan of intellectual property systems, it needs
it’s own set of IP laws that are specifically tailored to AI related IP.
In the next wave of research in the relationship between IP and AI
systems, the developments made by AI systems will require a stage of
“contextual adaptation” according to Launchbury (2017).
To better support the recognition of the work created by AI systems in
this stage and further the conversation on IP protection for AI works, this
would necessitate the AI system understand and perceive the real world
by themselves and no longer be reliant on massive data but learn more
from an understanding of the real-world phenomena and reason with it.
There are ambitions to create a wave of AI which allows machines and
humans to communicate more naturally. Launchbury further states that
the next wave of AI systems could create AI learning with the ability to
perceive complex and subtle information, conceive and learn within an
environment, execute the learning to create new meanings and exercise a
redaction process to make final decisions. The stumbling block for the
current wave of AI systems is an inability to satisfy step 3 of the copy-
right test.
With the next wave AI system, there is scope to reduce the need for
human intervention and to further expand the ability of the AI system
from a highly efficient tool to an AI system that can create new IP assets
using its own reasoning and equivalent thinking mind.
This is akin to the future success of AI systems that can not only start
thinking independently, but they are also capable of making internal
linkages within their digital environment, apply reasoning and deduction
skills to satisfy the human test component of intellectual property owner-
ship and resultingly, have an argument to benefit from the economic
protection that IP affords human owners. The door on IP protection in
broader areas for AI, other than computer generated works, is not
permanently closed. Rather, as the next wave of AI develops to mimic the

human mind and satisfy the tests in place for the different types of IP
assets, then it could open a door for AI systems to explore novel intel-
lectual property asset ownership and spearhead the creation of an AI/IP
protection system.
10 Legal and Ethical Considerations

Ethical discourse is commonly not a priority in a conventional study of
intellectual property and business. Moral sentiments often take a back
seat to market sentiments, even in shaping the direction of ethical busi-
ness strategy (Hoppe 2022). This anomaly persists despite growing inter-
est in ethics and IP. Taking an interdisciplinary and diverse perspective,
this section enriches the evolving definition and scope of ethical discourse
literature by focusing on intellectual property assets and regulation that
shape the business sector.
Considering the gap between theory and practice, and pertinent to the
legal and ethical considerations of a future AI/IP protection system,
bridging academic knowledge in unpacking ethical and governance issues
in the intellectual property industry.
11 Trustworthy AI
Although the advantages of AI in our daily lives are undeniable, there are
concerns about its dangers. Inadequate physical security, economic losses,
and ethical issues are just a few examples of the damage AI could cause.
As discussed earlier, the European Commission has identified examples
of unacceptable risk as uses of AI that manipulate human behavior and
systems that allow social-credit scoring (European Commission, 2021).
For example, the European legal framework would prohibit an AI system
similar to China’s social credit scoring (Marcia 2021). A legal framework
would require minimum level of uniformity at an international level, for
example, through the TRIPS agreement or a supplementary provision
with staggered grace periods dependent upon individual states’
18 N. Naim
socio-political conditions. Notwithstanding the need for specific criteria,

technical knowledge, certifications and implementation procedures
(Margoni 2022). The certifications are important as there needs to be an
independent governmental and international control mechanism on pro-
fessional standards for trustworthy AI.
12 Patent Offices and AI

One of the main areas where Patent Offices are considering the use of AI
is on how the role of AI in patent offices will be affected by the rapidly
developing frontier technologies underpinning the fourth industrial rev-
olution. IP Australia is developing a Trade Mark International
Classification Services (TMICS) to support customers with searching the
Madrid Goods and Service (MGS) database. Leveraging Natural
Language Processing (NLP) models (sentence-transformers), TMICS
assists with finding goods and services that match search terms and are
also semantically related. The aim is to significantly decrease the number
of queries and make the application process more efficient. It saves time
for the customers when they are searching through classifications for the
relevant goods and services, saving on effort and making the system more
user friendly. The attraction of using AI in IP offices is to aid the patent
offices in improving the quality of applications. This is also relevant to
making applications overseas and is one of the many AI initiatives that IP
offices are considering implementing (WIPO, 2018).
The UK IPO recently launched a consultation on potential legislative
changes for IPO digital transformation. Similar consultations are taking
place in the European Patent Office, the US Patent and Trademark Office
and national Patent Offices, to better understand AI-focused transforma-
tions to the IP protection and ownership process (Barazza 2021). The UK
IPO consultation concerned legal policy and regulatory frameworks, to
provide clearer guidelines and supporting mechanisms to address flaws
and gaps in the existing legal framework. The consultations are address-
ing the need to move away from paper processes and step into a digitally-
led era. The law was designed for a paper filing system and any proposed
changes to the law need to adapt to enable digital transformation of
services (UK IPO Consultation, 2023). Legislation could be altered to

better support innovation however this raises concerns over technical
expertise, risk management and robust monitoring of the legislative pow-
ers especially when delegated.
A preferred legislative approach is crucial to the success of the digital
transformation of patent offices as this will allow the patent office to
overhaul existing processes and provide a single, integrated digital system
to allow for a more efficient system for the registration of IP assets based
on whether it is a patent right, trademark or a registered design. As well
as building new digital services, governments need to ensure there are
legal framework that can support them. This requires flexibility to legally
adapt to new technologies, and to respond efficiently to then implement
the changes successfully. The economics of AI and AI-generated innova-
tion in IP is still in its infancy, the potential impact of international regu-
lation and how to resolve inconsistencies between varying approach taken
by other countries to the same regulatory issues, are part of the reasoning
behind giving more legislative powers to the IP offices. The complexities
arise when accounting for the rule of law and the lack of precedent for AI
friendly IP law.
13 AI in Copyright Law

Starting with the classification of AI in copyright law, both from a theo-
retical and AI governance point of view, there are a number of issues. In
the UK, copyright protection for computer generated works, without a
human author, exists and offers fifty years of protection (CDPA 1988).
To extend copyright protection beyond computer generated works, an
assessment of the current protection is valuable. In the UK IPO’s 2022
consultation, the copyright law for computer-generated works was found
to be effective, the AI had not been harmful and there had been no sig-
nificant change in the capacity level of AI systems to warrant further
protection, likewise, there was no marked negative impact to justify
reversing the current law.
Despite the many attractions of AI led IP law, most notably, for its
perceived economic benefits and ability to streamline current processes
20 N. Naim
and methods, the legislative changes require unbiased and independent

review. If there is to be a copyright protection system for AI, is it the AI
systems that self-regulate or is it the government through IP offices. AI
systems are built primarily by software designers and developers who are
not required to complete AI related professional ethics training. Thus, a
framework for AI Copyright law, which addresses the next wave of AI
and how it could meet copyright standards, also needs to deal with AI
regulation standards.
AI risks can be positioned against long-established responsible legal
principles of intention and culpability in civil and criminal law. AI sys-
tems lack capacity for subjective knowledge and intent, therefore proving
intent will require a new set of legal rules that set out the parameters for
establishing intent. (Yeung 2020). AI certification which evidences the
training of the software developer in core areas of ethics, intent, culpabil-
ity and negligence, would mitigate the risks posed by broadening AI
copyright law beyond computer generated works. An examination of
Public Law and the impact of AI on the general public, on future human
creators will need to be established as part of the overall framework to
ensure the AI is trustworthy and where, despite the infrastructure, it is
found not to be, requisite remedial protocols are in place.
For copyright law itself, approaching the place that AI has at the copy-
right dinner table is premature and ill-fitted. Even if the next wave or
subsequent waves of AI satisfy the copyright test, it is still a copyright test
for humans. Rather than shoehorn AI into a test that was never intended
for non-humans, an alternative test could be designed for AI that could
potentially grant copyright protection and remedies to the AI system,
both individually and jointly with a human. The duration periods will
require careful examination to reflect the speed at which AI can create
and whether, at least for the first cycle of AI protected copyright law,
there is a human check element. For text and data mining, the UK IPO
plan to introduce a new copyright and database exception which allows
text and data mining for any purpose. There will be safeguarding and
mechanisms to protect content, including a requirement for lawful access
(UK IPO Consultation, 2022).
14 AI in Patent Law

AI inventions and the possibility of patent ownership remains under
review at WIPO, the EU IPO, from the US and the UK. For AI-devised
inventions, although the UK IPO will review the legal provisions periodi-
cally, there is no change to UK patent law at the moment. Despite the
controversial Thaler case and the diverging views on the future of AI
based inventions, as discussed earlier, AI is not yet advanced enough to
invent without human intervention.
However, as can be seen from the speed at which AI is developing, the
margin between what AI can and cannot do without human interven-
tion, continues to shrink. The patent system supports AI innovation and
the use of AI as a tool in AI-assisted inventions, enabling human led or at
least involved inventions. Changing national laws will always prove to be
problematic as if there isn’t an international consensus on changing pat-
ent laws to give AI inventor rights, the level of protection is territorial and
minimal, especially if economic losses need to be recuperated. Contractual
clauses that can value the role of the AI system is an interim remedy to
recognise the full economic value of AI in patents until there is more
development in the AI systems to facilitate inventions without human
intervention at any stage.
15 Alas, Who Should the Intellectual

Property Belong To?
AI systems have developed rapidly and have been defined as AI-assisted
and AI-generated, and AI-assisted can be divided further into an earlier
and more premature stage of AI-supported. For AI-supported, it is the
first-wave of AI systems, embedded in rule-based programming with
human creators defining very specific and clear rules for AI systems to
follow, with little room for AI creativity. Turbo Tax falls within AI-assisted
as the AI systems do not have learning capability, nor are the systems
equipped to handle uncertainties and can only solve narrowly defined
problems.
22 N. Naim
Therefore, AI-supported systems are predictable by humans and there

is no innovation or creativity in the outcomes produced. Essentially,
AI-supported outputs are what the wave sys on the tin, they support the
production of an output by following rule-based programming defined
by humans, without demonstrating the components of intelligent ability.
AI-supported systems are devoid of the core human intelligent skills,
such as, learning, perceiving, abstracting and reasoning. It is clear that the
human engineers are the authors of AI-supported outputs and under the
legal rules of intellectual property, the rightful owners of the intellectual
property assets.
Where the conversation on intellectual property ownership gains trac-
tion is when considering the definitions of AI-assisted and AI-generated.
As a summary, the current law on this is clear. AI-assistance in the genera-
tion of IP assets is seen as a supportive tool in assisting the human cre-
ators and inventors. However, as highlighted in this chapter, and further
in subsequent chapters, the distinction between when there is human
involvement to when there isn’t needs to be reassessed and refined. The
clearer the distinction between AI assisted and AI generated, the more
incentive there is for AI systems to evolve to the next wave and meet the
current international IP standards. The extent and nature of human
involvement needs to be specified. If the AI-assistance is taking a lead in
the final production of the IP asset, the parameters of what is a minimum
level of human involvement also requires more transparency and clarity.
For AI-generated, the UK and other nations have exemptions in place, as
can be seen from the discussion of computer-generated works.
From preliminary analysis of the DABUS case and the Rembrandt 2.0,
the current legal provisions on IP protection have a short life expectancy
as there is strong evidence that AI systems are becoming more sophisti-
cated and the next wave of AI could see AI-generated outputs that fall
outside the scope of computer-generated works. The DABUS case is fur-
ther examined throughout the remaining chapters with diverging views
on its application and importance to patent law. The closer analysis of
patent and copyright law does support the existing EU Commission
views and the results of several consultations in the UK on whether AI
should have ownership rights in intellectual property law. In conclusion,
it is still too early to develop an IP system that has minimum level of
standards for AI. The reasoning for this is that the AI systems, albeit far
more sophisticated in the current wave of AI, still lack sentient ability and
despite the ability for AI to create a product, the role of a human is still
required. However, there will be a further wave of AI, and the exception
of computer-generated works will not suffice to meet the intellectual
effort of AI. The next stage for intellectual property law policy and ethics
is to create the framework where AI and humans can both have an equi-
table slice of the IP pie.
References
Barazza, S. The Year Of Data?, Journal of Intellectual Property Law & Practice,
Volume 16, Issue 2, February 2021, Pages 89–90, https://doi.org/10.1093/
jiplp/jpab024
Bulayenko, O, Frosio, G and Geiger, C. Opinion of the CEIPI on the European
Commission’s Proposal to Reform Copyright Limitations and Exceptions in
the European Union, Centre for International Intellectual Property Studies
(CEIPI) Research Paper No. 2017–09.
Cochrane, D and Mhangwane, C. Dabus: rise of inventive machines, (2022).
Spoor and Fisher. Available at https://spoor.com/dabus-the-rise-of-
the-inventive-machines/.
Cooke, C. Will Page says “music copyright has never had it so good” as global
revenues reach almost $40 billion (2022), Complete Music Update.
Copyright, Design and Patent Act 1988
EU Commission on Civil Law Rules on Robotics (2015/2103(INL)
European Commission, Coordinated Plan On Artificial Intelligence 2021 Review.
Available at https://digital-strategy.ec.europa.eu/en/library/coordinated-plan-
artificial-intelligence-2021-review
Fogel, D. Evolutionary Computation: Toward a New Philosophy of Machine
Intelligence, Third Edition, (2005), Wiley.
Grauwe, D P and Gryspeerdt, S. Artificial intelligence (AI): The qualification of
AI creations as “works” under EU copyright law, (2022), Grewer.
Guadamuz, A. Artificial intelligence and copyright (2017) World Intellectual
Property Office Magazine.
Hoppe, M. Theories and Applications of Artificial Intelligence. (2022). United
Kingdom: States Academic Press.
24 N. Naim
Infopaq International A/S v Danske Dagblades Forening (C-5/08)

EU:C:2009:465 (16 July 2009).
Launchbury, J. ‘DARPA Perspective on AI’ (Defense Advanced Research Projects
Agency) 2017. https://www.darpa.mil/about-us/darpa-perspective-on-ai
Li, L. Intervention Report for the WIPO Conversation on Intellectual Property
and Artificial Intelligence (Third Session) (2020) WIPO https://www.wipo.
int/export/sites/www/about-ip/en/artificial_intelligence/conversation_ip_ai/
pdf/ind_li.pdf
Minsky, M. Steps Toward Artificial Intelligence in Proceedings of the IRE, Vol.
49, No. 1, 1961, pp. 8–30
Marcia, V. The EU path towards regulation on artificial intelligence,
Techtank, 2021.
Margoni, T. A Deeper Look into the EU Text and Data Mining Exceptions:
Harmonisation, Data Ownership, and the Future of Technology, GRUR
International, 71(8), 2022, 685–701.
Rodgers, W. Artificial Intelligence in a Throughput Model: Some Major
Algorithms (Routledge 2020)
Thaler v Comptroller General of Patents Trade Marks And Designs [2021]
EWCA Civ 1374 UK Intellectual Property Office Consultations and
Publications
U.S. Copyright Office, Report To The Librarian Of Congress By The Register
Of Copyrights, 1966
Walter Thompson Amsterdam, 2016 World Intellectual Property Office
Yeung, K. Response to European Commission White Paper on Artificial
Intelligence. 2020.
Fostering Innovation by Utilising Big
Data: The Data Act and the Risk
of Quasi-Exclusivity Reinforcing Data
Lockups
Marc A. Stuhldreier
1 Introduction
In Thaler v Commissioner of Patents [2021] FCA 879, a judge of the
Federal Court of Australia for the first time found that an artificial intel-
ligence (AI) can satisfy the requirements to be regarded as an inventor
according to Australian patent laws. While the judgement was revoked in
2022, and other courts around the world do not seem to follow the initial
court’s approach anytime soon, this case provides an indication of how AI
and the Big Data which feeds its learning processes slowly but steadily
tend to pick up a ‘life’ of their own. Smart devices, as part of the Internet
of Things (IoT), that utilise AI have become prevalent in almost any
aspect of our day to day lives, including smart homes, the agricultural
sector, and medicinal products. The development, appropriate use and
advancement of such smart technologies requires vast amounts of data,
M. A. Stuhldreier (*)
Linköping University, Linköping, Sweden
e-mail: marc.stuhldreier@liu.se

https://doi.org/10.1007/978-3-031-42576-9_2
26 M. A. Stuhldreier
commonly referred to as Big Data. Thus, the collecting and processing of

such data becomes increasingly important to modern life, and the con-
temporary smart way of living. The mining of such data constitutes an
essential requirement for the smart technologies business sector and is
pivotal for its future.
This requirement is starkly contrasted, however, by the rights of those
who generate the data. Users of smart technologies become workers for
the creation of value to industry, even in their leisure-time, irrespective of
their awareness of this fact (Noto La Diega 2022a). Notably, the gener-
ated user-data may regularly be of a personal nature wherefore this way of
collecting data constitutes much more than simple work. Further, ques-
tions arise as to the ownership of user generated data which currently is
under the factual control of the data holder, with users having few oppor-
tunities to access and utilise this data themselves. This leads to a problem
of so-called data enclosures or data lock-ins, by which data holders retain
the power over the user generated data and lock this data away, prevent-
ing not only the re-use of this data by the users, but also hinder access to
such data for innovative purposes. Considering the legal and ethical ques-
tions concerning the generation and utilisation of this type of data, the
traditional legal distinction between online-offline issues is not sufficient
to deal with the intricacies of big data, AI, and smart technologies (Noto
La Diega 2022a). Various specialised laws and regulations were thus
introduced, regulating specific issues in specific fields. Adding to this
wide array of regulations, on 23 February 2022 the EU Commission
proposed the Regulation on Harmonised Rules on Fair Access to and Use
of Data (Data Act) which was formally endorsed by the European
Parliament on 9 November 2023 and adopted by the Council on 15
November 2023. As a key pillar of Europe’s strategy for dealing with data,
the Data Act shall become an important contribution to adequately facil-
itating digital transformation. The main scope of the Data Act regulates
the conditions under which and by whom value can be created from data.
To this end, the Data Act adopts specific objectives, including (1) setting
up of rules for the utilisation of user-generated data as to ensure fairness
in its use, (2) establishing consistency relating to data access rights, which
under current (non-)regulation may vary distinctly, and (3) increasing
the amount of available data, benefiting companies, citizens, and public
administrations alike. In addition to these main objectives, the EU
Fostering Innovation by Utilising Big Data: The Data Act… 27
recognises the need of facilitating access to data for the purpose of foster-
ing innovation with societal objectives, and for responding to current
challenges including in the fields of public health, agriculture and for
achieving the Sustainable Development Goals (SDGs). With its 50
Articles, the Data Act thus sets ambitious goals as to creating a level play-
ing field for the various stakeholders in what may be called user-generated
data. The task of striking an adequate balance is complicated by the num-
ber of contrasting interests at stake. Concerns can be raised over ques-
tions of access to data, its ownership, the protection of personality rights
of the users that provide the data, and particularly their right to self-
determination. On the other hand, increasing access to user-generated
data calls into question the protection of the legitimate rights of the
inventors of smart technologies. Ultimately, while recent EU legislation
aims for opening data lockups to make data accessible and re-usable, the
scope of these regulations seems to in fact strengthen the position of data
holders, providing them with further legal measures to maintain their
data enclosures (Noto La Diega 2022b).
In this respect, this contribution scrutinises the provisions of the Data
Act to analyse whether the Data Act is successful in striking an adequate
balance between the rights and obligations of the various stakeholders in user-
generated big data. Further, this chapter seeks to ascertain whether the
Data Act provides for sufficient accessibility to facilitate that data can be re-
used for innovation activity toward societal objectives.
2 Utilising the Value of Big Data:

Innovation in the Fourth Industrial
Revolution (4IR)
2.1 Big Data
In the digital age, data can be loosely described as a unit containing digi-
tal information. While information was always a valuable asset to human-
ity and thus sensitive for businesses, states and citizens, data can now be
defined as a building block in the modern digital economy (Burri 2021;
Irion 2021). This digital economy can also be described as data-driven in

that it is “built around the collection, preservation, protection, imple-
mentation and understanding of many different types of data.” (Aaronson
2021). The McKinsey Global Institute suggested in a 2016 report that
the global economy has now entered an era that is ‘defined by data flows’
(Manyyika et al. 2016). Indeed, in today’s economy, data, or better big
data, has introduced a new era in global trade, in which data has transi-
tioned into becoming a commodity in itself (Gervais 2021).
Big data refers to the amassing and storing of extensive and exponen-
tially growing data sets. While there is no single generally valid definition,
big data is often defined as datasets exhibiting three essential features, a
fourth additional feature and a derivative fifth feature. These features can
be conceptualised as the five V’s; high volume, veracity, velocity, variety,
and value. Volume refers to the size of data sets or simply the sheer
amount of data generated. Veracity of automatically collected data refers
to data quality and its trustworthiness. Velocity concerns the pace at
which data is generated, while variety refers to the different data types
and its numerous sources. Lastly, value refers to the current or future
value of these datasets, for corporations and other stakeholders with
regard to their utilisation for economic, innovative, or societal objectives
(Burri 2021; Gervais 2021). In addition, big data can be generally cate-
gorised as either structured or unstructured. Structured data is data that
is sufficiently clear to be utilised for a purpose, while unstructured data,
which may include meta data collected as a side-product of using a digital
device or different types of free texts in online forms, generally requires
further structuring before its value can be unlocked (Jülicher 2018).
Today, exponential amounts of data are generated by machines and
human activity in the digital society, where smart devices and smart infra-
structure, including smart homes, smart health, and smart cities, are
readily available and frequently used in modern life (Irion 2021; Storr
and Storr 2017). In fact, data-collection opportunities arise in most
aspects of modern life where almost everything is connected by data-
driven technologies and mobile internet (Aaronson 2021). In the sphere
of the Internet of things (IoT), the success of wearable technologies, or
wearables, exemplifies how data can be collected and to a certain extent
analysed in real time when valuable health information is gathered by
fitness trackers and successively processed in their connected apps. While

the large amounts of data generated by wearables can be utilised for the
user’s and society’s benefit, the risk of this data being abused for creating
user profiles, providing unwanted insights into their personal sphere rises
simultaneously. This is particularly concerning when considering that
users of smart technologies predominantly lack the knowledge to under-
stand how their data is processed and utilised (Jülicher and Delisle 2018).
Notably, as types of body-attached computers, wearables not only
come in the form of fitness trackers, but also as smartwatches and smart
glasses which generate enormous amounts of user specific data, covering
most aspects of day-to-day life. Innovative wearable technologies are fre-
quently developed with researchers working, among others, on smart
contact lenses, smart tattoos, and smart fashion, such as intelligent socks,
providing further opportunities for the production and analysis of expan-
sive datasets (Jülicher and Delisle 2018; Storr and Storr 2017). This digi-
talisation of society and the economy, which is now also regarded as the
fourth industrial revolution (4IR), leads to a growing fusion of different
technologies, “rewriting the boundaries between biological, physical, and
digital.” (Noto La Diega and Derclaye 2023).
The utilisation of big data is generally based on linking the data that
was collected from different devices and other sources. Oftentimes, data
is initially collected without a specific purpose and stored until it may
become useful at a later time. Linking in the big data context is thus not
only concerned with combining data that was collected from different
sources but also with combining new data and existing prior generated
data. Considering that a large amount of such data is generated by users
through their use of personal smart devices, these datasets often contain
personal data, which, according to the purpose limitation principle of the
General Data Protection Regulation (GDPR), can legally only be col-
lected for specific purposes (Culik 2018). While not all user generated
data necessarily is of personal nature, big data and the mode of its collec-
tion often blur the lines between personal and non-personal. In fact, user
generated data can rarely be fully anonymised (Burri 2021).
To derive value from big data, and particularly for unlocking the
potential of unstructured data, relevant information needs to be extracted
from large datasets and organised. Such activities are commonly referred
to as data mining. Notably, data mining can be conducted for purposes

that differ from the purposes for which the data was initially collected
(Jülicher 2018). As a manual evaluation of the sheer amount of data col-
lected in big datasets is virtually impossible, the mining of data is con-
ducted by utilising AI algorithms that facilitate machine learning
processes, enabling the machines to create high value outputs and gain
new insights from the raw data (Gervais 2021; Jülicher 2018). These
insights then not only reflect the past, but can actively be utilised to pre-
dict future trends (Jülicher 2018).
2.2 The Importance of Big Data for Future

Innovation in the 4IR
The importance of big data for the digital economy lies in its potential
value not just for the original data collector, but for a variety of stakehold-
ers (Jülicher 2018). Additional stakeholders include the users of smart
technology as well as, for example, manufacturers, aftermarket service
providers, insurances, health service providers, and public and private
sector researchers (Jülicher and Delisle 2018). Fully unlocking the poten-
tial of big data can provide enterprises with previously unknown infor-
mation. New insights derived through data analytics, such as identified
patterns and conclusions that indicate future trends, are big data’s pre-
dominant economic value, helping businesses to improve their decisions
and strategies (Storr and Storr 2017; Seuba 2021). Unsurprisingly, big
data is now of enormous economic interest, both for its value as a busi-
ness asset as well as for its future value that can be derived by unlocking
its innovative potential. In this regard, big data, and particularly personal
data, is often referred to as the new oil (Storr and Storr 2017). By not
only facilitating more efficient business operations but by also fostering
better innovation big data has vast potential to positively serve the public
interest (Burri 2021).
It is recognised today that data is key to achieving sustainability and
that adequately unlocking its potential is of enormous value for the com-
mon good. Access to and utilisation of data, for example, are considered
to be a key factor in achieving climate justice and sustainable innovation
(Noto La Diega and Derclaye 2023). Similarly, health data collected from
wearables can be utilised to improve personal health by amending per-
sonal habits before health conditions occur, thereby helping to mitigate
modern public health issues (Storr and Storr, 2017). To make adequate
use of the information contained in big datasets, and to achieve these
societal objectives, it is crucial to unlock the potential of the collected
data. To be effective, it is important to recognise that data mining can be
conducted to gain new insights from currently unused big data, not only
by the data collector but by other stakeholders with access to the data as
well. For this to be effective at large, however, open access to the locked-
up data would be required. While competitors on the markets and the
general public thus have both economic and public interests in gaining
access to and utilising available data, the collectors of big data and those
who analyse it tend to also seek (legal) protection for their respective
interests in the data and its value (Gervais 2021). Overly strong protec-
tion, either via legal instruments such as intellectual property (IP) laws,
or through the factual control exercised by the data holder, can in turn
lockup data, hamper its utilisation, and prevent the realisation of big
data’s full potential for the societal benefit.
2.3 Unlocking the Value/Potential of Big Data
One of the main concerns is that currently a relatively small number of

companies dominate the global digital economy and thereby exercise a
quasi-control of the digital markets, both in developed markets as well as
in the developing world (Gervais 2021). This control of the digital econ-
omy was facilitated by a modern phenomenon termed ‘digital disposses-
sion’ by which data collectors amass vast amounts of personal and
non-personal data through appropriation. Digital dispossession further
provides data holders with a new type of ‘data power’ which facilitates
further data lockups with detrimental impacts on both innovation and
aftermarket services (Noto La Diega and Derclaye 2023). When consid-
ering the importance of data in modern power dynamics, questions arise
as to how data can be utilised appropriately, how data ownership should
be regulated, how access can be facilitated, and how incentives can be
maintained for investments into the generation of new data.
2.3.1 Data Ownership
As a baseline, it can be emphasized that Data is frequently generated by

human action, with machines being merely used for its processing (Noto
La Diega and Derclaye 2023). Concerning the ownership of user gener-
ated data, it is relatively straight forward to suggest that actual content
created by a user, such as email texts, are clearly the ownership of the user.
Similarly, photographs taken by a user with the camera of a connected
devices fall within the IP of that user, the author of that image. However,
the situation may be less clear with further data that is collected through
the camera, via sensors or through user input in devices (Storr and Storr
2017). In general, users tend to lack having any influence regarding the
use and evaluation of such data (Jülicher and Delisle 2018). Particularly
when personal data is processed, technology allows the collectors or hold-
ers of user generated data to create user profiles supplemented with infor-
mation about location, habits, interests, political views or the sexuality of
the users (Röttgen 2018, von Schönfeld and Wehkamp 2018). If appro-
priated, email texts, for example, can be analysed for key words and facili-
tate further profiling opportunities with economic benefits for the data
collector (Röttgen 2018). The prevalent power asymmetries between data
collectors and technology users generally leave the users with almost no
individual control over the processing of their data (Thouvenin and
Tamò-Larrieux 2021).
Narrowing down the perspective, there exist two main issues concern-
ing the ownership of data that need to be balanced: (1) the rights of users,
including the protection of their privacy in relation to personal data, and
(2) the rights of data holders with respect to investment protection and
intellectual property. The concept of data ownership can then be under-
stood two-fold as a property right in data, either directly introduced by
law or indirect through a position of factual control which is protected by
law (Thouvenin and Tamò- Larrieux 2021). It is to note that currently
there exists no IP right or similar property right for data (yet) (Storr and
Storr 2017; Podszun and Pfeifer 2022). Without such a right, and in
consideration of the enormous interest in data by a variety of stakehold-
ers, questions arise as to who should be the owner of the data, or whose
interests take precedent (Storr and Storr 2017). In reality, while there is
no general legal property right to data, data holders can rely on other
measures such as trade secret protection and database rights as well as
technical and contractual measures to effectively establish a property-like
protection, providing them with a technical-factual position of data own-
ership (Noto la Diega 2023; Podszun and Pfeifer 2022; Storr and Storr
2017). Additionally, in the EU, the creator of a databank can own the
data contained in this databank via a sui generis right established by the
EU Database Directive (96/9/EC).
The main concern with property and property-like rights is that his-
torically, property rights derived from the need to regulate scarce
resources. Data, however, is not scarce and its use by one party does not
exclude another party from extracting value from the same data (Storr
and Storr 2017; Hornung and Schomberg 2022). This general non-
scarcity of data is also where its main potential lies if data is made suffi-
ciently accessible (Hornung and Schomberg 2022). With various
technical difficulties already hampering the adequate sharing of data, a
property-like right in data, adding a further layer of complications, can
severely exacerbate problems concerning the accessibility and utilisation
of data by the various stakeholders. Ultimately, property rights tend to be
of an exclusive nature, meaning that non-right holders, including the
general society the very users who generate the data, can be excluded
from enjoying the benefits arising from big data (Storr and Storr 2017).
While this recognition provides a strong argument against a property
right to data, a balanced approach to the question of ownership also
needs to take account of concomitant questions and debates concerning
the protection of investments and the maintaining of incentives for the
future generation of data. In essence, similar to the debates on industrial
IP rights, the incentive argument suggests that investors in data generat-
ing products and services need to be incentivised to make such invest-
ments by receiving protection which prevents competitors who did not
make similar investments from using the generated data (Kerber 2022).
Conversely, it may be argued that such incentives are less relevant in the
context of big data, particularly when collected via IoT devices, as the
data collection occurs automatically as a by-product of the regular use of
such devices. Thus, the required investment into the data collection itself
is marginal, and the development of the IoT devices is likely to be already

protected by conventional IP rights (Storr and Storr 2017). Additionally,
the value of the data itself, even if simultaneously used by other stake-
holders, may provide a sufficient incentive for companies to invest into
its collection. As recognised by the EU Data Act, the situation may be
different for small and medium sized enterprises (SMEs), who may in
fact face substantial investments to set up a data collection business. New
start-ups could thus potentially be disincentivised if their collected data
would be open to everyone (SWD(2022) 34 final Annex 8).
All things considered, however, and with particular regard to the vast
amounts of data that are continuously collected, it seems unsuitable to
suggest that further legal incentives for the collection of data are required
(Noto La Diega and Derclaye 2023). In this respect, the Max Planck
Institute for Innovation and Competition concluded in 2016 that there
is neither a requirement nor a justification for introducing a legal exclu-
sive right to data (Drexl et al. 2016). Quite the opposite, it seems that
new incentives are actually required to open up data as data holding
enterprises oftentimes refrain from voluntarily sharing and facilitating
the utilisation of their data for innovation by others (Noto La Diega and
Derclaye 2023).
The discussion so far focussed on the concept of ownership for data
collectors. From a different perspective, data ownership, or a property
right to data could be granted to the users of connected devices, who,
through their use, create the data in the first place. It is questionable,
however, whether such a property right of the user would be fruitful, or
whether it would, in fact, harm the interests of users. While on the sur-
face, a property right would provide the user with the power over their
data, a property right can also be traded away, with potential detrimental
impacts on the user’s privacy. Once sold, the data property to personal
data could be owned by corporations, which would potentially negate the
positive protection impacts of privacy laws, such as the GDPR (Storr and
Storr 2017; Thouvenin and Tamò-Larrieux 2021). One of the concomi-
tant risks is that the buyer of a user’s (personal) data could successively
exclude the user from accessing and using their own personal data
(Thouvenin and Tamò-Larrieux 2021). However, at least with regard to
the right to erasure, the GDPR seems to override any other agreements
made between the individual and other (economic) actors. It thus seems
that, compared to an ownership right in user generated non-personal
data, privacy rights are stronger in that they cannot generally be traded
away (Storr and Storr 2017). In theory, data ownership could thus poten-
tially help mitigating some of the prevailing problems of the digital econ-
omy by providing individuals with a better negotiation position vis-à-vis
other economic actors (Thouvenin and Tamò-Larrieux 2021). Then
again, due to prevalent information asymmetries, it would be easy for
corporations to gain ownership of the data, for example, through hidden
contractual terms, a problem that already can be observed in the field of
data privacy where the use of intransparent consent tick boxes is
commonplace.
Ultimately, as long as clear regulations on data ownership and access
rights are missing, data collectors and developers of connected devices
can implement technical designs for products that provide them with the
de-facto control over the data collected by their devices, which
then enables them to effectively exclude competition and to prevent other
enterprises and organisations from using the available data for innovative
activities (Kerber 2022).
2.3.2 A Right to Access
In addition to questions of data ownership, the second big issue for data
governance in the digital economy concerns the accessibility and re-use
of data. As of yet, there are no general data access rights, neither for busi-
nesses vis-à-vis other businesses, nor for government vis-à-vis businesses
(Thouvenin and Tamò-Larrieux 2021). The situation, however, is about
to change, with the implementation of the European Data Strategy, inter
alia, via the EU Data Governance Act, already providing for the accessi-
bility of publicly held data for further utilisation and re-use, and other
regulations, including the Data Act, following in the near future. From a
business perspective, data is generally viewed as a valuable asset that
should not be freely shared, particularly with competitors (Thouvenin
and Tamò-Larrieux 2021). Access to data, however, is a precondition for
market entry, participation in a supply chain and innovation (SWD(2022)
34 final at 11). Demands of fairness, free competition, and the facilita-

tion of innovativeness thus all require the establishment of better data
accessibility (Podszun and Pfeifer 2022).
Quasi-exclusive rights to data, protecting investments, on the other
hand, are liable to abuse and can be utilised to restrict competition with
negative impacts on market growth and innovation, further hindering
technological development (Storr and Storr 2017). As a result, restric-
tions of the accessibility of data are likely to have adverse impacts on the
digital economy at large (Thouvenin and Tamò-Larrieux 2021). In fact,
even before the age of digital data, barriers to the free flow of information
were considered detrimental to economic and innovative activity (Burri
2021). If regarded as a public good, the full potential of data could be
realised through its utilisation by a variety of stakeholders, with negligible
detrimental impacts on any individual actor (Thouvenin and Tamò-
Larrieux 2021).
However, to establish data as a public good, it is key to implement a
balanced approach to data access that promotes trust in the digital econ-
omy. To enhance this trust, boundaries to the free flow of data must be
defined, at least balancing access rights with privacy concerns. In addi-
tion to the protection of personal data, certain essential business interests,
including trade secrets and other IP rights, should be taken into consid-
eration as well (Thouvenin and Tamò-Larrieux 2021; Noto La Diega and
Derclaye 2023). To be effective, this balancing of rights should re-consider
traditional conceptualisations of IP, particularly re-evaluating the extent
to which IP is actually conducive to innovation and identifying where the
over protection of private interests starts to hamper technological prog-
ress for the public benefit. Ultimately, facilitating access to data alone will
not suffice in fostering innovation if this data cannot legally be utilised
(Noto La Diega and Derclaye 2023).
In addition to a general accessibility, unlocking the full potential of big
data further requires the free flow of data across borders (Ferracane 2021).
This recognition, as well as concomitant obstacles, were emphasized by a
Japanese government initiative on ‘Data Free Flow with Trust’ at the
2019 G20 summit. In this respect, the G20 Osaka Leaders’ Declaration
summarises: ‘Innovation is an important driver for economic growth,
which can also contribute to advancing towards the SDGs and enhancing
inclusiveness. […] Cross-border flow of data, information, ideas and

knowledge generates higher productivity, greater innovation, and
improved sustainable development, while raising challenges related to
privacy, data protection, intellectual property rights, and security’
(G20 2019). Similar to data lockups at private corporations, the problem
in a global context is that governments tend to aim for restrictions of data
flows to other countries, as data is regarded as a valuable asset for their
domestic industries. Such restrictions of data flows then are liable to
hamper innovativeness on a global scale and thus impede the develop-
ment of products for the public benefit (Burri 2021, Ferracane 2021).
On the other hand, particularly when personal data is at stake, the miti-
gation of legitimate privacy concerns may in fact require certain limita-
tions on permissible data transfers and processing activity in other
countries to protect the fundamental rights of domestic data subjects
(Burri 2021).
Nonetheless, the UN High-Level Panel on Digital Cooperation
emphasized in 2019 that in a global economy that is increasingly reliant
on digital interdependence, ‘new forms of digital cooperation’ are required
‘to ensure that digital technologies are built on a foundation of respect for
human rights and provide meaningful opportunity for all people and
nations’ (United Nations 2019 at 6). In a similar vein, the OECD recog-
nises the responsibility of both governments and other stakeholders to
facilitate a common digital economy that ‘improves peoples’ live[s] and
boost[s] economic growth for countries at all levels of development, while
ensuring that nobody is left behind’ (OECD 2018). The free-flow of data
across borders is consequently regarded as a highly relevant factor for a
fair distribution of value and for facilitating location-independent inno-
vation that serves the public interest globally, potentially contributing to
the effective realisation of human rights. In this respect, open data strate-
gies are increasingly implemented and demanded at an international
stage by countries of the Global South, which seek to protect their legiti-
mate interests in the technology driven world (Noto La Diega and
Derclaye 2023).
Lastly, for the unlocking of big data to be effective, and to enhance its
accessibility, it is key to tackle the technical factual power of data holders.
In this respect, questions arise as to the quality of data as well as to the
portability and interoperability of data sets. Interoperability issues for

example arise when technology producers adopt proprietary systems for
their data collection and processing that effectively prevent others from
utilising the data (Jülicher and Delisle 2018). Facilitating adequate access
to data thus requires both the introduction of legal measures that define
and regulate questions concerning the ownership of and access rights to
data as well as the implementation of means that challenge protectionist
technical measures that are liable to thwart the free flow of data.
3 The EU Data Strategy and the Data Act

3.1 The European Data Strategy
Recognising the high value of big data, and the challenges surrounding
its utilisation, the EU adopted the Digital Single Market (DSM) Strategy,
intending to unlock the economic and social potential of the digital
economy. One of the DSM’s key aims lies in facilitating the free flow of
data to improve its accessibility and utilisation, and to facilitate innova-
tion in the EU (Thouvenin and Tamò-Larrieux 2021; Hennemann and
Steinrötter 2022). According to the German Ethics Council it is crucial
to recognise the value of data as a social resource (Hornung und
Schomberg 2022). One of the key challenges of the digital transforma-
tion is to facilitate adequate data access for both individuals and busi-
nesses in accordance with fair competition standards, while simultaneously
safeguarding the protection of privacy and consumer rights as well as of
IP rights. To round-up the potential for value generation from big data,
the EU aims to likewise facilitate the transfer of knowledge and data
between public and private data holders. Free flow of data is further
regarded as a key component of European competitiveness in the global
digital economy (Thouvenin and Tamò-Larrieux 2021).
In February 2020, the EU supplemented the DSM by adopting the
European Data Strategy which aims at creating the necessary precondi-
tions for a genuine European data economy to establish Europe as a
global leader in the digital economy. The Data Act is one of the key pillars
of this strategy (COM(2022) 68 final; Specht-Riemenschneider 2022a;

Podszun and Pfeifer 2022). As a result of the Covid-19 pandemic, the
recovery plan of the EU further recognises that its data strategy needs to
establish a data economy that, inter alia, facilitates innovation
(SWD(2022) 34 final). In addition to the Data Act, the DSM Strategy
and the European Data Strategy rely on a number of legal policies and
regulations, including the General Data Protection Regulation (GDPR),
the Data Governance Act, the Digital Markets Act, the Digital Services
Act, the Artificial Intelligence Act as well as sector specific regulations
such as the proposed European Health Data Space (EHDS).
The GDPR provides regulations for the protection of personal data
and the fundamental right to privacy of European Citizens, by establish-
ing limitations to the permissible processing of personal data (Regulation
(EU) 2016/679). At the core of the Data Governance Act, on the other
hand, lie regulations for facilitating the further utilisation and re-use of
data held by public sector bodies, for facilitating the mutual utilisation of
data between private and public actors, and for furthering data altruism
(Hornung and Schomberg 2022). The Data Governance Act further
strives to create the necessary preconditions for individuals and corpora-
tions to voluntarily share their data without jeopardising their existing
rights over this data (COM(2022) 68 final). In particular, the Data
Governance Act permits the reuse of certain types of public sector data,
even where such data may be commercially confidential or protected by
IP rights. This permission, however, is balanced through mechanisms
safeguarding the protection of IP, as protected data is strictly shared on
the basis of confidentiality (Noto la Diega 2023).
The Digital Markets Act regulates the behaviour of the biggest digital
enterprises, i.e. those corporations that provide core platform services
while enjoying a “durable and entrenched position”, significantly impact-
ing the European internal market, aiming to prevent them from being
gatekeepers of the internet. In comparison to the other regulations of
the European digital economy, it is notable that the Digital Markets Act
has a strong focus on accessibility while not making any commitments
towards the protection of IP (Noto la Diega 2023). The Digital Services
Act, on the other hand, promotes responsibilities of private actors towards
preventing the abuse of digital services for conducting illegal activities

online (Noto la Diega 2023).
While not generally prohibiting high-risk AI systems, the European
Artificial Intelligence Act, provides mandatory regulations for their utili-
sation, requiring transparency to facilitate the appropriate interpretation
and use of their output. The main aim of this regulation is to protect the
fundamental rights of European citizens in a world where the economy
increasingly relies on AI systems that impact technology users, with
potentially detrimental consequences (Noto la Diega, 2023).
Lastly, the proposed European Health Data Space provides sector spe-
cific regulations for the adequate utilisation of health data. The EDHS is
the first proposed sector specific regulation to supplement the Data Act,
subject to the Data Acts horizontal regulations (COM(2022) 197 final).
As can be conferred from the amount of different legal policies for the
regulation of the European digital economy, one of the core problems of
the European Data Strategy is that these different legislative measures
regulate different, but also almost all aspects of the data economy.
Potentially conflicting targets of these legislative measures, however, are
not sufficiently taken into consideration (Veil 2022). The Data Act for
example supplements both the Data Governance Act and the GDPR
without sufficiently balancing their interrelation with each other (Specht-
Riemenschneider 2022a).
3.2 The Data Act
In recognition of the importance of data for the digital economy and for
achieving the green and digital transition as well as the fact that data
remains under-utilised where its value is held locked-up by a small num-
ber of large corporations which hampers the realisation of the full poten-
tial of the digital age, on 23 February 2022, the EU introduced a Proposal
for a Regulation on harmonised and fair access to and use of data (Data
Act, or DA) which aims to establish fairer access to and a fairer value
allocation from non-personal data. The proposed Regulation was adopted
by the Council on 15 November 2023. Unlocking enclosed data provides
immense potential for strengthening a sustainable data economy in
Europe that can reduce the “digital divide” and shall create beneficial
opportunities that are available to everyone. To achieve this, data access
and its use need to be regulated accordingly. Notably, the European
Council explicitly concluded in October 2021 that the regulatory frame-
work for facilitating this accessibility of data shall be conducive to inno-
vation (COM(2022) 68 final at 1). As will be elaborated below, however,
the Data Act is unlikely to achieve its goal of creating an environment
that enables the full potential of innovativeness as exclusive rights and
quasi-exclusive rights that protect data holders remain in place and con-
tinue to restrict the accessibility of data for research purposes. The prob-
lem of insufficient data access and use for societal purposes is thus not
sufficiently mitigated by this regulation.
In March 2021, the European parliament further recognised and high-
lighted the need for the creation of European data spaces that should
enable the free flow of data between different sectors and between public
and private stakeholders, including academia (COM(2022) 68 final at
2). While this statement found its way into the Explanatory Memorandum
of the Data Act, the Act itself provides no direct data access rights for
academia, limiting any public sector access to situations of exceptional
need, such as public emergencies (Arts. 14 and 21 DA) or to situations
where data access is granted to academia as a third party by request of the
user of a device (Art. 5 DA). The European Parliament further high-
lighted the importance of tackling restrictions of competition, barriers to
market entry and broader issues concerning the access and use of data
that arise from data enclosures. In this regard, the Data Act aims at facili-
tating that EU businesses of all sectors receive opportunities through
which they can be innovative and competitive (COM(2022) 68 final at 2).
Importantly, the Explanatory Memorandum clarifies that the provi-
sions on data access do not change the scope of previous regulation and
protection of data through IP rights, which remain unaltered with the
notable exception of the sui generis right established by the Database
Directive. Similarly, the Data Act will not amend previous European data
legislation on sectoral level, but any (future) sectoral legislation should
“be aligned with the horizontal principles of the Data Act” (COM(2022)
68 final at 5). In this regard, the Data Act was proposed as a horizontal
full-harmonisation regulation that shall provide basic rules applicable to
all sectors, so that future sectoral regulation is subject to the general

requirements of the Data Act (SWD(2022) 34 final at 7). Therefore, it is
to be expected that the Data Act will set the standard for data access
rights for the foreseeable future (Specht-Riemenschneider 2022c).
In general, the Data Act pursues the achievement of five targets: (1)
facilitating access to and the use of data by consumers and businesses
without impeding on incentives for the generation of new data, (2) facili-
tating the utilisation of data by public sector bodies in situations of
exceptional need, (3) facilitating the switching between cloud and edge
service providers, (4) providing safeguards against unlawful data transfers
by cloud service providers, and (5) providing for the adoption of interop-
erability standards for data to enhance accessibility (COM(2022) 68 final
at 3; From Noto La Diega and Derclaye 2023). The overall objective of
the Data Act then lies in maximising the value generation from data for
both the economy and society by granting data access to a wider range of
stakeholders, facilitating its further use, while balancing these access
rights with incentives for the continued collection of new data. Among
the specific objectives, the EU further holds that incentives for the shar-
ing of data should facilitate that other businesses can proactively engage
in the data economy, which is regarded as a precondition for fostering
innovation (SWD(2022) 34 final at 26 and 27).
The now 50 Articles of the Data Act are divided into 11 chapters.
Chapter I lays down general provisions, including the scope and subject
matter of the data act as well as relevant definitions. Chapters II, III and
IV regulate the accessibility and utilisation of data for private actors,
including the data holder, users of connected devices and third parties.
Chapter V then specifically regulates the access to privately held data by
public sector bodies. The regulations provided by Chaps. VI, VII and
VIII address specific issues concerning the portability and interoperabil-
ity of data as well as safeguards for when non-personal data is shared in
an international context. While Chaps. IX and XI concern the imple-
mentation and enforcement of the Data Act, Chap. X clarifies the rela-
tion between the Data Act and the Database Directive by providing
limitations to the applicability of the sui generis right to database protec-
tion. The Data Act shall establish preconditions for facilitating access to
data which is generated through the use of connected devices and
services. Central to the regulations of the Data Act is thus the creation of
mandatory data access rights for both private and business users of con-
nected devices (Specht-Riemenschneider, 2022a). Importantly, the Data
Act shall facilitate innovation by opening data enclosures, while main-
taining sufficient incentives for the generation of new data (Hennemann
and Steinrötter 2022; Podszun and Pfeifer 2022). To achieve the aim of
opening data enclosures, the Data Act regulates data accessibility by
addressing three main socio-economic dimensions: business to consumer
(B2C) relationships, business to business (B2B) relationships, and
Business to Government (B2G) relationships (Podszun and Pfeifer 2022).
According to Article 1, the Data Act applies to:
(a) manufacturers of connected products placed on the market in the

Union and providers of related services, irrespective of the place of
establishment of those manufacturers and providers;
(b) users in the Union of connected products or related services;
(c) data holders, irrespective of their place of establishment, that make
data available to data recipients in the Union;
(d) data recipients in the Union to whom data are made available;
(e) public sector bodies, the Commission, the European Central Bank
and Union bodies that request data holders to make data available
where there is an exceptional need for those data for the performance
of a specific task carried out in the public interest and to the data
holders that provide those data in response to such request; (f ) pro-
viders of data processing services, irrespective of their place of estab-
lishment, providing such services to customers in the Union;
(g) participants in data spaces and vendors of applications using
smart contracts and persons whose trade, business or profession
involves the deployment of smart contracts for others in the context
of executing an agreement (Art. 1(3) DA).
Notably, similar to the applicability of the GDPR, the Data Act fol-
lows the lex loci solutionis principle, meaning that its applicability is sub-
ject to the relevant marketplace. In other words, the regulations of the
Data Act apply to all data holders when data is generated within the
European market (Hennemann and Steinrötter 2022). Additionally,
Article 1(5) DA provides that the Data Act is without prejudice to the
applicability of EU data privacy and data protection laws, and that the
access rights of Chap. II of the Data Act shall supplement the access and
portability rights of the GDPR. Considering, however, that user gener-
ated data in many instances is likely to be of personal nature, the Data
Act leaves various tensions between data access and data protection largely
unresolved (Kerber 2022).
Article 2 DA provides definitions that apply for the purposes of the
regulation. Here, data is defined as “any digital representation of acts,
facts or information and any compilation of such acts, facts or informa-
tion, including in the form of sound, visual or audio-visual recording”
(Art. 2(1) DA). The term data is thus not simply regarded as synonymous
to the term information but is rather regarded as a means for transporting
information (Hennemann and Steinrötter 2022). Notably, storage
devices and mere online services are a priori excluded from the scope of
the Data Act (Specht-Riemenschneider 2022b). Furthermore, recital 15
specifically excludes derivative data – i.e. data that results from the pro-
cessing or analysis of raw data – from the regulations of the Data Act
(Recital 15 DA). This exclusion is particularly regrettable as raw data can
often be too limited to be re-usable, while the re-use of derivative data
could provide valuable insights, benefitting research for societal purposes
(Podszun and Pfeifer 2022). In contrast to raw data, however, the cre-
ation of derivative data requires an effort by the data holder and may thus
be protected by IP rights. Where such data is not protected by IP rights,
the Data Act may then strengthen the exclusive position of data holders
in the future. For achieving the aim of promoting innovation towards
sustainability and the common good, however, it is key that adequate
data accessibility is designed in a way that facilitates the unlocking of the
full value of big data (Noto La Diega and Derclaye 2023).
3.2.1 B2B and B2C Data Access
According to Article 3 DA, products are to be designed in a way that

facilitates easy direct product data and related service data accessibility for
their users (accessibility by design), provided that such direct access is
secure and appropriate (Art. 3 DA). However, a mere direct accessibility

of data may not be sufficient in that the regulation leaves unclear whether
providing in-situ access can suffice to fulfil this accessibility requirement
(Podszun and Pfeifer 2022). When read in conjunction with Article 4
DA and recitals 8 and 22, it seems data holders can fulfil their obligations
by simply providing online data access without being required to share a
copy of the relevant data (Kerber 2022). While in-situ access may seem
like an adequate safeguard to mitigate the risks of unnecessary data trans-
fers, this limitation can also provide an easy pretext for data holders to
circumvent the actual sharing of their data (Kerber 2022; Podszun and
Pfeifer 2022). In addition to the accessibility by design requirement,
Article 3(2) and (3) DA introduce a transparency requirement that shall
protect the interests of users by enabling them to make informed deci-
sions on the processing of their generated data before they enter into a
contract with the data holder (Hennemann and Steinrötter 2022).
Articles 4 and 5 DA provide users of connected devices with non-
waivable rights that facilitate the accessibility and sharing of their gener-
ated data for all legal purposes, subject to certain limitations (Kerber
2022). While Article 3 DA establishes the accessibility by design require-
ment, Article 4 provides both private and business users of connected
devices with a direct right against the producer to access and use their
generated data, if the producer is also the data holder (Art. 4(1) DA). The
aim of this right is to make data accessible to the users whose usage of a
device is a key component for the production of this data. In relation to
the GDPR, it is important to note that according to recital 7, the data
access right of Article 4 DA does not create a new legal basis permitting
the processing of personal data (Hennemann and Steinrötter 2022). The
control of users is further strengthened by Article 4(13) DA, providing
that the data holder may only process user generated data on a contrac-
tual agreement with the user (Art. 4(13) DA). At a first glance, this may
be regarded as an attribution of the data ownership to the user, as the data
holder can only realise the data’s economic value with the user’s consent
(Hennemann and Steinrötter 2022). In reality, however, this hurdle is
easy to overcome for data holders as distinct guidelines for consumer
protection applicable to relevant contractual terms are missing. This
problem can already be observed with respect to informed consent under
the regulation of the GDPR, where, due to information asymmetries,

data subjects often struggle to understand what they are consenting to
exactly. Similarly, the current text of the Data Act does not prevent the
data holder from implementing buy-out contracts by which the user con-
sent becomes an integral part of the purchase or leasing contract of a
device (Specht- Riemenschneider 2022a, b; Podszun and Pfeifer 2022).
Taking account of potential conflicts between data access rights and IP
rights, Article 4(6) DA clarifies that data access cannot be denied for
reasons of trade secrecy, but that instead measures need to put in place
that sufficiently protect trade secrets, for example through the conclusion
of an NDA (Art. 4(6) DA; Hennemann and Steinrötter 2022). However,
Article 4(8) DA provides that in exceptional circumstances where the
trade secret holder can “demonstrate the high probability of suffering
serious economic damage from the disclosure of trade secrets, despite the
technical and organisational measures taken by the user”, data access by
the user can be refused on a case-by-case basis (Art. 4(8) DA). To protect
the legitimate interests of data holders, Article 4(10) DA prevents users
from utilising the accessed data for the development of competing prod-
ucts (Art. 4(10) DA). While this prohibition may restrict activities that
could lead to the development of improved products and enhance com-
petition, it also provides a reasonable limitation to access rights so that
the initial developers of innovative products are not disincentivised from
being innovative (Podszun and Pfeifer 2022).
In accordance with recital 5, Article 5 DA facilitates that users cannot
only access and use their data themselves, but further share this data with
third parties or request that a third party receives direct access to the user
data from the data holder (Art. 5 DA, Recital 5 DA). However, the Data
Act does not establish a general direct access right for third parties
(Specht-Riemenschneider 2022a). Data access by third parties is thus
dependent on the willingness of the user to grant such access (Hennemann
and Steinrötter 2022). While this restriction of third-party access is ade-
quate from a consumer protection perspective, it also severely limits the
facilitation of the re-use of data for research purposes (Specht-
Riemenschneider 2022b). Similar to Article 4(6), Article 5(9) DA pro-
vides safeguards for trade secrets of the data holder. In general, it can be
observed that the Data Act prioritises access rights over trade secret
protection, as long as sufficient measures are taken to preserve their con-

fidentiality (Specht-Riemenschneider 2022c). With respect to third-party
data transfers, however, the data holder is only required to disclose trade
secrets “to the extent that such disclosure is strictly necessary to fulfil the
purpose agreed between the user and the third party” (Art. 5(9) DA).
This prioritisation of data access over trade secrets is later somewhat
mitigated by Art 8(6) DA which provides that “an obligation to make
data available to a data recipient shall not oblige the disclosure of trade
secrets” unless specifically provided for by Union law or the national
implementation of Union law (Art. 8(6) DA). Article 11 DA then pro-
vides safeguards for data holders, enabling them to implement technical
measures such as smart contracts and encryption to ensure that the shared
data is not accessed or used in an unlawful manner by third parties (Art.
11(1) DA).
Such protection is particularly important for safeguarding the confi-
dentiality of trade secrets, as even where conventional protection for trade
secrets is in place, its enforceability can be jeopardised once a secret has
been made public. Smart contracts mitigate this risk by enabling the data
holder to gain insights into if and how the shared data was used by third
parties (Bartke et al. 2022). Article 11(2) and (3) DA provide further
defensive rights for data holders against the illegitimate utilisation of
their data which strengthens their de-facto exclusive position. This may
potentially lead to an over protection of the data (Specht-Riemenschneider
2022b). While implementing technical safeguards is generally a reason-
able measure for protecting the legitimate interests of data holders, a
potential abuse of smart contracts, however, may be liable to hamper the
utilisation of shared data for innovation purposes.
Strengthening the position of the user toward third parties, Article 6
DA provides that third parties shall only process the data they receive
under Article 5 subject to purposes and conditions agreed upon with the
user (Art. 6(1) DA). The user decides thus not only if and with whom
they want to share their generated data, but also what the data can be
used for by third parties. While this is again a welcome inclusion of con-
sumer protection measures, a purpose limitation for the use of non-
personal data can adversely impact innovativeness, as research may
discover new, previously unknown purposes for utilising such data to
achieve societal objectives. Similar to the limitation of the usage rights of

users, Art 6(2)(e) further restricts innovative activity by preventing third
parties from using data received via the user for developing products that
are in competition with the products and services of the data holder from
which the data originates (Art. 6(2)(e) DA). Additionally, the purpose
limitation is supplemented with further safeguards, including a prohibi-
tion for the third party to share the user data with another third party, or
better a fourth party, unless this is necessary for the delivery of services
commissioned by the user (Art. 6(2)(c) DA). To protect free competition
on the European Single Market, Article 6(2)(d) DA specifically prevents
users from sharing data they receive from the data holder with third par-
ties who, according to the Digital Markets Act, are designated as gate-
keepers to prevent those gatekeepers from further strengthening their
market positions by gaining data from competitors via their users (Art.
6(2)(d) DA).
Even when a user decides to share their generated data with a third
party, to utilise this data the third party is required to enter into two con-
tractual relationships both with the user and with the data holder. In
other words, a user cannot directly share their data with the third party
without a licensing agreement between the third party and the data
holder. A third party who intends to use data for research and innovation
may also require access to aggregated datasets. The requirement to have
contractual relations with the data holder and each individual user, how-
ever, makes acquiring aggregated datasets not just difficult, but poten-
tially impossible. Therefore, the regulations for third party access in the
Data Act fail in establishing necessary preconditions for facilitating
meaningful innovation (Kerber 2022).
To prevent data holders from abusing their factual control over the
data and to mitigate power imbalances in contract negotiations, Article
8(1) DA provides that data shall be made accessible to users and third
parties on fair, reasonable and non-discriminatory (FRAND) terms (Art.
8(1) DA; Hennemann and Steinrötter 2022). While according to Article
9, the data holder can require a reasonable payment for data transfers, the
data holder cannot freely determine the fees and conditions for data
access by third parties (Kerber 2022). For small and medium-sized enter-
prises (SMEs) and non-profit organisations fees are even limited to the
actual costs incurred by making the data available (Art. 9(4) DA). In
general, the Data Act seems to emphasize the specific needs of small
enterprises, as Article 7 DA completely exempts data holders who qualify
as small and micro enterprises according to Union law from the B2C and
B2B data sharing requirements, provided that they are not linked to
other enterprises or partner organisations (Art. 7 DA).
3.2.2 B2G Data Access
In addition to the B2C and B2B data access rights, chapter V of the Data
Act provides a direct data access entitlement for the public sector, i.e.
public sector bodies or Union Institutions, agencies and bodies. The
main regulation facilitating public sector access, Article 14 DA, however,
strictly limits the access right to situations of exceptional need (Art. 14
DA). Article 15 DA then defines that an exceptional need to use data
shall be limited in time and scope, and only exists in circumstances:
(a) where the data requested is necessary to respond to a public emer-

gency and the public sector body, the Commission, the European
Central Bank or the Union body is unable to obtain such data by
alternative means in a timely and effective manner under equivalent
conditions;
(b) in circumstances not covered by point (a) and only insofar as non-
personal data is concerned, where:
(i) a public sector body, the Commission, the European Central

Bank or a Union body is acting on the basis of Union or national
law and has identified specific data, the lack of which prevents it
from fulfilling a specific task carried out in the public interest,
that has been explicitly provided for by law, such as the produc-
tion of official statistics or the mitigation of or recovery from a
public emergency; and
(ii) the public sector body, the Commission, the European Central
Bank or the Union body has exhausted all other means at its
disposal to obtain such data, including purchase of non-per-
sonal data on the market by offering market rates, or by relying
on existing obligations to make data available or the adoption of

new legislative measures which could guarantee the timely avail-
ability of the data (Art. 15 DA).
Article 17 DA provides further administrative requirements for data

requests by public bodies. The restrictive requirements of Articles 15 and
17 seem to be designed to make data access by governments a rare excep-
tion (Specht-Riemenschneider, 2022c). And even where governments
gain access to user generated data, the utilisation of such data is subject
to further strict limitations under Article 19 DA, including a purpose
limitation, so that shared data can only be used to respond to the situa-
tion of exceptional need and must be destroyed once it is no longer
required for achieving that purpose (Art. 19(1) DA).
While Article 21 DA facilitates that data accessed by public sector
bodies can be further shared with non-profit organisations that conduct
scientific research with a public interest mission, this right is similarly
limited to the purposes for which the data was initially requested, mean-
ing the research can only be conducted to respond to the specific excep-
tional circumstances (Art. 21 DA). One of the aims of granting public
sector access is to make it easier for public sector bodies to obtain and
re-use privately held data for public interest purposes (SWD(2022) 34
final Annex 3). A limitation to situations of exceptional need is thus too
narrow. In this regard, it is particularly regrettable that there is no direct
access right for public research institutions outside of situations of excep-
tional need. To adequately unlock data enclosures, thereby facilitating
meaningful and swift innovation, access to data is required on situation-
independent basis. On the other hand, limitations for government access
to data may nonetheless be important to mitigate the risk that govern-
ments obtain a general opportunity to evaluate and analyse product data
in an abusive manner to the detriment of their citizens (Podszun and
Pfeifer, 2022).
3.2.3 International Data Transfers, Interoperability

Standards, and the Sui Generis Right
Article 32 DA provides regulations and safeguards concerning interna-

tional access to and transfer of data. In particular, these regulations aim
to protect data held within the EU from unlawful access by international
actors and governments if such access would be in conflict with Union
law or the law of EU Member States (Art. 32(1) DA). The regulations
provide an additional layer of protection for non-personal data, as inter-
national transfers of personal data are already regulated by the
GDPR. While, if adequately applied, the regulations of Article 32 protect
the rights and interests of European data holders and users of connected
devices, international data flows are also an important driver for innova-
tion and a requirement for a fair value allocation in the global digital
economy. It is therefore crucial that any restrictions to the free flow of
data are sufficiently balanced.
To unlock data enclosures and to facilitate the adequate accessibility of
data, it is important that data can be utilised by the recipients of shared
data without the requirement of substantial additional investments. In
this respect, Article 33 DA introduces essential requirements that opera-
tors of data spaces need to fulfil to facilitate the interoperability of their
data (Art. 33 DA). These requirements should make it easier for actors
other than the data holder to find relevant data and reduce transaction
costs for when the data is shared and re-used (SWD(2022) 34 final at
27). Thereby, the sharing of different types of data between different pro-
viders shall be made possible, which, in turn, shall facilitate the creation
of common data spaces in strategic sectors for the public benefit (Podszun
and Pfeifer 2022).
Lastly, the Data Act recognises at least some of the potentially detri-
mental impacts of IP rights on the sharing of data. In this regard, Article
43 DA explicitly excludes the applicability of the sui generis right of
Article 7 of the Database Directive for data obtained from the use of con-
nected devices to prevent the Database Directive from jeopardising the
effectiveness of the Data Act (Art. 43 DA).
3.2.4 Initial Critique of the Data Act
All things considered, there seems to be a consensus among scholars and

commentators that the Data Act is unlikely to achieve its goals. In par-
ticular, it is suggested that the Data Act, as well as the other EU policies
on the digital economy, fail in providing sufficient preconditions for
facilitating open access to data (Noto La Diega and Derclaye 2023). The
Commission Staff Impact Assessment Report accompanying the Data
Act suggests that “the Data Act would enable wider data use across the
economy, notably by regulating the fundamental questions of who can
use the data generated by connected products and related services, and
what are the conditions for such use.” (SWD(2022) 34 final) As this is
supposed to facilitate consumer choice and promote innovation, it is
regrettable that considerations of public research are not sufficiently
included in the legislation. A limitation, or better exception, allowing
business-to-government data access, including for public research pur-
poses, only in exceptional situations is simply too narrow to foster any
meaningful innovation for the public benefit. Recognising that in accor-
dance with the Data Governance Act, public data is made available for
re-use by private corporations, it seems reasonable to suggest that a simi-
lar general access right for the public to utilise data held by enterprises for
societal objectives should be considered (Thouvenin and Tamò-Larrieux
2021; Noto La Diega and Derclaye 2023).
Furthermore, the Data Act, and the European Data Strategy more
generally, do not conclusively resolve conflicts between openness require-
ments and property rights. While in essence, through the Data Act, the
EU rejects the idea of a de jure property right to data, new safeguards are
implemented that consolidate property-like positions of data holders
(Hennemann and Steinrötter 2022; Kerber 2022; Noto la Diega 2023;
Podszun and Pfeifer 2022). Thereby, the Data Act effectively strengthens
the technical-factual control of the data holder. Open data access thus
only exists on paper but is negated by the various limitations of the Data
Act (Veil 2022; Specht-Riemenschneider 2022a, b). The technical-factual
control of the data holder is thus the general rule supported by the Data
Act, which is only restricted in the exceptional circumstances that a user
utilises their data access rights (Specht-Riemenschneider, 2022c). As the

data holder remains in control over the data, it seems that any access is an
exception which requires a justification. Conversely, it should be ques-
tioned whether it is not the quasi-exclusive control of the data holder
that requires a justification in the first place (Specht-Riemenschneider
2022a, b). As long as the law accepts the technical-factual control of the
data holder as the default position, the Data Act cannot achieve its aim of
unlocking large amounts of data for innovation purposes. The access and
sharing rights are too limited and thus too weak while the effective con-
trol of the data holder is strengthened (Kerber 2022). In particular, data
holders gain the potential to delay or even prevent data access by third
parties, with long negotiations and potential disputes hampering swift
access for innovation purposes (Podszun and Pfeifer 2022).
All things considered, the Data Act’s acceptance of the de facto control
of data holders strengthens their position with legal entitlements which
are similar to the exclusive rights provided by IP protection, making
access an exception to the norm. The Data Act thus urgently requires re-
balancing as a legally accepted de-facto control provides data holders
with a quasi-exclusive right to data, which in turn may consolidate poten-
tial monopolistic market positions (Kerber 2022).
4 A de-Facto Exclusive Right to Data

By enabling the creation of monopolies or monopolistic market positions
for rights holders, the granting of exclusive rights constitutes a restriction
of the free competition principle. In the intellectual property debate, and
in consideration of the social contract theory, the restriction of competi-
tion through exclusive rights is arguably justified by the incentives they
provide for innovativeness with potential public benefits. Investors are
incentivised by receiving an opportunity to profit from risky investments,
or at least a chance to recover their R&D investments (Bartke et al.
2022). The granting of exclusive rights under IP laws is not uncontrover-
sial and was heavily criticised over the past decades. In particular, it can
be suggested that in vital sectors, such as health, the public benefits of
innovation are negated by exclusive rights that facilitate monopolistic
market positions and lead to higher prices. To be justified, it is thus cru-

cial to ensure that exclusive rights are adequately balanced as any overpro-
tection would negatively tip the balance between the legitimate interests
of inventors or investors and the interests of society at large in favour of
the former.
Additionally, balanced IP regulations, or the regulation of non-scarce
resources more generally, need to be adapted to current economic and
scientific conditions. Historically, when IP laws were first introduced,
innovation was considered a closed process, meaning that the innovative
process was entirely conducted by a single inventor or at least within the
confinements of a single enterprise. However, the approach to innovation
has changed over the past decades and an opening of the innovative pro-
cess, by which corporations increasingly rely on external knowledge, can
be observed. This trend, suitably termed open innovation, reflects the
increasing reliance of innovators on cooperation rather than competition,
which enhances and accelerates innovation processes for the benefit of all
stakeholders (Bartke et al. 2022). The transition towards open innovation
calls into question whether traditional IP systems and their reliance on
the provision of exclusive rights remain fit for their purpose in an open
innovation environment. The experience with Covid-19 and the acceler-
ated development of vaccines highlights the crucial importance of open
innovation, as access to health and research data was key for responding
to the pandemic in a timely manner, particularly when compared to the
regular duration required for the development of new medicines. On the
other hand, industry may claim that in an open innovation context, there
is actually a growing need for protection of the interests of those who
invest in data generation, to incentives the sharing of such data by miti-
gating detrimental impacts on investors and by clarifying current legal
uncertainties (Bartke et al. 2022).
Similar to the experience with IP rights, the de-facto control over data
can impede on free market dynamics. When a data holder has the control
over an extensive data set generated by their products and can exponen-
tially increase this “data treasure” by exclusively utilising the data to
develop improved products and thereby surpass all competition, the data
holder may end up in a monopolistic position on the market. This is
detrimental to the market entry and innovativeness of competitors. An
adequate stimulation of innovativeness would thus require adequate

access to this data by third parties (Bartke et al. 2022). Broad access rights
to such data would be comparable to regulations for standard essential
patents (SEPs), access to which is required even where the patent holder
is unwilling to voluntarily grant a license. To tackle this problem, SEP
holders are now required to grand licenses to everyone on FRAND terms
(Bartke et al. 2022). While the Data Act specifically includes a reference
to FRAND terms in Article 8, in contrast to SEPs, there still exists no
general data access right for third parties. The limitations provided by the
Data Act rather strengthen the position of data holders vis-à-vis competi-
tors who require data access, consolidating their exclusive position.
4.1 The Problem with Exclusivity: Lessons

from the Global Harmonisation of Intellectual
Property Rights
As the arguments for a de facto control of data holders are quite similar
to the debates concerning patent rights, the following analysis shall exem-
plify the conflict between exclusive rights and innovation by reference to
a notable experience that emerged from the international harmonisation
of minimum patent protection standards through the WTO TRIPS
Agreement. One of the main arguments for introducing patent rights for
all fields of technology, including the pharmaceutical sector, in countries,
and particularly developing countries that formerly did not provide for
the patentability of pharmaceuticals was that patents provide inventors
with an incentive for being innovative. The premise of stronger global
patent protection was then to increase encouragement for the develop-
ment of new medicines, including products that serve the interests of the
developing world. Particularly in the field of pharmaceuticals, the devel-
opment of new products involves substantially high costs and are regarded
as a high-risk investment (Abbott 2015; ECOSOC 2001; Hestermeyer
2007; Reichman 2009). Granting exclusive rights via the provision of
patents should then incentivise high risk investments by facilitating the
successful commercialisation of new medicines through the provision of
time-limited monopoly positions on the markets (Abbott 2011; Flynn
et al. 2009). As tackling global disease burdens not only requires the
affordability of existing medications, but also the future availability of
new medicines, it can be argued that limited exclusive rights provide a
balance between the interests of industry and the wider societal interests
(Abbott 2015; Henry and Searles 2012; Phillips 2006).
While it may be true that an adequate level of patent protection can be
beneficial for stimulating innovation, overly strong or excessive protec-
tion may in fact have detrimental impacts on innovation by creating new
obstacles for future research activity. As innovation is frequently the result
of research activity that builds upon prior discoveries, a stringent patent
protection of older inventions is liable to prevent researchers from utilis-
ing existing knowledge, thereby delaying future innovation (Abbott
2015; Richards 2008; Smith et al. 2009). A continuous proliferation of
patent protection therefore directly contradicts its very purpose of stimu-
lating innovation by restricting the innovative activity of competitors.
The exclusive rights granted under patent protection thus can be used to
effectively hinder innovation (Richards 2008; ECOSOC 2001). This
concern is particularly prevalent in the pharmaceutical sector, where the
development of an administrable medicine is the result of several stages
of innovation. As each of these stages can potentially suffice for fulfilling
patentability requirements, the granting of patents can create further
obstacles for researchers who cannot utilise the results of previous devel-
opment steps if a competitor receives exclusive protection (ECOSOC
2001). By restricting opportunities for utilising existing knowledge, the
exclusive rights granted to patentees are therefore liable to ultimately
delay technological progress (Richards 2008). By and large, it can be
summarised that the global introduction of pharmaceutical patent rights
created new obstacles that hamper appropriate research efforts, thereby
defeating the very purposes for which they were introduced. Adding a
new layer to this protection by consolidating quasi-exclusive rights for
data holders of user generated raw data, particularly from smart health
devices, would in all probability create further obstacles, aggravating the
already problematic situation.
4.2 An Exclusive Right to Data and the Problem

of IP Overlaps
In the modern digital economy, the introduction of a quasi-exclusive

right to data would further considerably strengthen the defensive posi-
tion and strategy of rights holders, who can rely on their data rights in a
worrying synergy with other IP rights. This synergy lies in the potential
overlapping of exclusive protection. Just as in the traditional regulation of
IP rights, the Data Act effectively provides for access rights, and particu-
larly public and third-party access, as an exception to the general rule.
Notably, each regulation of the different types of IP and data rights has
its own permissibility requirements for granting exceptions. This becomes
particularly problematic, where different exclusive rights cover the same
substance. In today’s economy, there exists a “growing fusion between
hardware, software, digital content, services, data and biotechnologies.”
(Noto La Diega and Derclaye 2023). A single product may, for example,
enjoy protection of various patents, trademarks, trade secrets and now,
potentially include a de facto control of the data. The exceptions to each
of these rights are quite specific, and similar exceptions for the same
exceptional reasons may not be enshrined in the regulation of another
right. An exception for public health reasons under patent laws, for
example, may not exist in the same manner under trademark law or trade
secret protection. When multi-IP protection exists, the exceptions of
each individual set of rights may thus become insufficient when the rights
holder can rely on another right to maintain their protection. While a
legitimate exception under patent rights would facilitate access, the non-
existence of an equivalent exception could render such access nevertheless
an infringement of trademark or trade secrecy laws (Noto La Diega and
Derclaye 2023; Noto La Diega 2022a). An exclusive rights system that
subjects access rights to specific exceptions thus faces the risk that other
exclusive rights may remain in place, restricting adequate and legitimate
accessibility. Ultimately, the more stringent exclusive protection systems
are liable to prevail, thereby diminishing the efficacy of exceptions pro-
vided by systems that are more generous towards access (Noto La Diega
and Derclaye 2023).
As the Explanatory Memorandum to the Data Act provides that IP

rights remain unaltered, the Data Act leaves potential conflicts between
IP and data accessibility unresolved, or better excludes it from its focus
(COM(2022) 68 final at 5; Noto La Diega and Derclaye 2023). By sug-
gesting that the Data Act is without prejudice to IP rights and trade secret
protection the regulation aims to guarantee the continued protection of
rights holders (SWD(2022) 34 final Annex 8). While it is sensible to sug-
gest that the legal rights of IP owners generally should be protected, in
respect to data access, it is also clear that IP rights impact innovation in
different ways, and not just by providing incentives for investments.
When aiming to unlock data for innovation purposes, existing problems
with IP protection standards should therefore be taken into deeper con-
sideration. Attempting to regulate data access while ignoring the pre-
existing issues arising from stringent IP protection is simply insufficient.
Thus, while the general premise of enhancing data access through the
Data Act is good, IP needs to be accounted for to enable meaningful
innovation. The categorical exclusion of IP considerations from the Data
Act does nothing to mitigate current legal uncertainties. Rather, it creates
new uncertainties where the laws and their teleological purposes stand in
conflict. Accordingly, the consolidation of a quasi-exclusive position of
data holders is likely to aggravate the already existing problems of exclu-
sive IP rights, where access is merely side-lined to the exceptions.
5 Conclusion
This chapter has highlighted the importance of big data for the modern
digital economy and particularly the crucial importance of facilitating
access to data to enhance meaningful innovation toward societal objec-
tives. Section 3 has introduced the European strategy on how data can be
used more efficiently, including consideration of how access for the vari-
ous stakeholders shall be facilitated. In this respect, this contribution
took a deeper look at the initial proposal of the Data Act that shall estab-
lish fairer access to and a fairer value allocation from non-personal data.
The unlocking of data enclosures shall thereby strengthen a sustainable
data economy in Europe, creating beneficial opportunities for everyone.
Notably, the accessibility and utilisation of data shall be regulated in a

manner that is conducive to innovation. As elaborated in this contribu-
tion, however, the Data Act is unlikely to achieve its objectives.
It is welcome that the Data Act introduces safeguards that prevent
gatekeepers from profiting from data access rights and utilising their
power to gain access to further data, while at the same time recognising
the specific needs of small and micro enterprises and start-ups (Podszun
and Pfeifer 2022). Similarly commendable is that the Data Act intro-
duces interoperability requirements to facilitate the creation of common
data spaces to enhance the re-usability of data. To re-use data in the first
place, however, stakeholders require adequate access rights. This is where
the Data Act is likely to fail. In the outset, this contributions questions
whether the Data Act is successful in striking an adequate balance between
the rights and obligations of the various stakeholders. Upon close exami-
nation, it can be observed that while the Data Act aims at facilitating the
access and re-use of data, the specific regulations that grant access rights
are limited by a number of conditions. While it may be relatively easy for
the user of connected decides to gain access to their created data, it is
likely that a mere in-situ access suffices to fulfil the requirements of the
Data Act. As the user then would not receive a copy of their data, the re-
use of such data would be hampered. Data access is further complicated
for third parties who have an interest in utilising the data for innovation
purposes or for the provision of after market services. Access for third
parties is dependent on the willingness of either the data holder or the
specific user. A general access right for third parties is non-existent. If
access is granted by a user, the third party is further required to enter
contractual agreements both with the data holder and with each indi-
vidual user, which makes the acquisition of aggregated datasets for third
parties almost impossible. Such aggregated datasets, however, would be
required for fostering meaningful innovation. For the data collector, on
the other hand, the limitations to the access rights in the Data Act imply
that even though the EU rejects the concept of a legal property right to
data, the Data Act strengthens the de facto control of the data holder.
Thereby, the Data Act grants the data holder a quasi-exclusive position.
This leads us to the second question raised by this chapter of whether
the Data Act provides for sufficient accessibility of data to facilitate its
re-use for innovation activity towards societal objectives. The answer to

this question likewise is negative. The quasi-exclusive de facto control,
and the fact that third parties have no direct access right to data, provides
data holders with the opportunity to prevent or delay access to such data
for research purposes. This position is particularly strong, where a prod-
uct resulting from research activity may stand in competition with the
product from which the raw data originates. More important from a pub-
lic interest perspective, however, is that the Data Act does not establish a
general access right for academia or other public research institutions.
Any access by the public sector is limited to situations of exceptional
need. This is regrettable, particularly when considering that the Data
Governance Act makes publicly held data available for commercial
exploitation by private enterprises. A similar access right for the public to
utilise privately held data for societal objectives therefore would be desir-
able. All in all, it seems that the Data Act trips over the same pitfalls as
modern international IP protection standards when it comes to regulat-
ing access rights. So, while the Data Act is an urgently required valuable
step in the right direction, to be adequate for regulating the modern digi-
tal economy, its provisions require further re-balancing. In particular, the
limitations to access rights should be re-calibrated to ensure that data
accessibility provides a proportionate counterbalance to the de facto con-
trol of data holders to prevent the introduction of a legally accepted
quasi-exclusive right to data.
References
Aaronson, S. A. (2021). Data is Different, so Policymakers should pay Close
Attention to its Governance. In M. Burri (Ed.), Big Data and Global Trade
Law (pp. 340–359) Cambridge University Press
Abbott, F.M. (2011). Intellectual Property and Public Health: Meeting the
Challenge of Sustainability. Global Health Programme Working Paper
No7/2011. Retrieved December 12, 2023, from https://papers.ssrn.com/
sol3/papers.cfm?abstract_id=1965458
Abbott, F.M. (2015). Trade in Medicines. In R. Smith et al. (eds), Trade and
Health: Towards Building a National Strategy (pp. 117–136) World Health
Organization
Bartke, L. et al. (2022). Der Schutz von Trainingsdaten de lege ferenda: What
would Machlup do? RDi Recht Digital, 431–439
Burri, M. (2021). Data Flows and Global Trade Law. In M. Burri (Ed.), Big
Data and Global Trade Law (pp. 11–41) Cambridge University Press
Culik, N. (2018). Brussels Calling: Big Data and Privacy. In T. Hoeren, &
B. Kolany-Raiser (Eds.), Big Data in Context: Legal, Social and Technological
Insights (pp. 29–35) Springer Nature
Drexl, J. et al. (2016). Data Ownership and Access to Data – Position Statement
of the Max Planck Institute for Innovation and Competition of 16 August
2016 on the Current European Debate. Max Planck Institute for Innovation
& Competition Research Paper No. 16–10. Retrieved December 12, 2023,
from http://dx.doi.org/10.2139/ssrn.2833165
Ferracane, M. F. (2021). The Cost of Data Protectionism. In M. Burri (Ed.), Big
Data and Global Trade Law (pp. 63–82) Cambridge University Press
Flynn, S., Hollis A. and Palmedo M. (2009). An Economic Justification for
Open Access to Essential Medicine Patents in Developing Countries. 37
Journal of Law, Medicine & Ethics, 184–208
G20, (2019, June 29). G20 Osaka Leaders’ Declaration. Retrieved December 12,
2023, from https://www.consilium.europa.eu/media/40124/final_g20_
osaka_leaders_declaration.pdf
Gervais, D. J. (2021). TRIPS Meets Big Data. In M. Burri (Ed.), Big Data and
Global Trade Law (pp. 160–176) Cambridge University Press
Hennemann, M. & Steinrötter, B. (2022). Data Act – Fundament des neuen
EU-Datenwirtschaftsrechts? NJW Neue Jusristische Wochenschrift, 1481–1486
Henry, D. and Searles, A. (2012). Chapter 9: Pharmaceutical Pricing Policy. In
Management Sciences for Health (ed.) MDS-3: Managing Access to Medicines
and Health Technologies (at 9.2–9.22) 3rd edn, Management Sciences
for Health
Hestermeyer, H. (2007) Human Rights and the WTO: The Case of Patents and
Access to Medicines. Oxford University Press
Hornung, G. and Schomberg, S. (2022). Datensouveränität im Spannungsfeld
zwischen Datenschutz und Datenzugang: das Beispiel des Data Governance
Acts. Computer und Recht, 508–516
Irion, K. (2021). Panta Rhei: A European Perspective on Ensuring a High Level
of Protection of Human Rights in a World in Which Everything Flows. In
M. Burri (Ed.), Big Data and Global Trade Law (pp. 231–242) Cambridge
University Press
Jülicher, T. (2018). Education 2.0: Learning Analytics, Educational Data

Mining and Co. In T. Hoeren, & B. Kolany-Raiser (Eds.), Big Data in
Context: Legal, Social and Technological Insights (pp. 47–53) Springer Nature
Jülicher, T. and Delisle M. (2018). Step into „The Circle“ – A Close Look at
Wearables and Quantified Self. In T. Hoeren, & B. Kolany-Raiser (Eds.), Big
Data in Context: Legal, Social and Technological Insights (pp. 81–91)
Springer Nature
Kerber, W. (2022). Governance of IoT Data: Why the EU Data Act Will not
Fulfill Its Objectives. GRUR International, 72(2), 120–135
Manyyika, J. et al. (2016). Digital Globalisation: The New Era of Global Flows.
McKinsey Global Institute
Noto La Diega, G. (2022a). Internet of Things and the Law: Legal Strategies for
Consumer-Centric Smart Technologies. Routledge
Noto La Diega, G. (2022b, November 04). New chapter! “Ending Smart Data
Enclosures: The European Approach to the Regulation of the Internet of Things
between Access and Intellectual Property. Blog post retrieved December 12,
2023, from https://guidonotoladiega.wordpress.com/2022/11/04/new-
chapter-ending-smart-data-enclosures-the-european-approach-to-the-regu-
lation-of-the-internet-of-things-between-access-and-intellectual-property/
Noto La Diega, G. (2023). Ending Smart Data Enclosures: The European
Approach to the Regulation of the Internet of Things between Access and
Intellectual Property. In S.-A. Elvy and N. Kim (Eds.), The Cambridge
Handbook on Emerging Issues at the Intersection of Commercial Law and
Technology (Forthcoming) Cambridge University Press
Noto La Diega, G. and Derclaye, E. (2023). Opening Up Big Data for
Sustainability: What Role for Database Rights in the Fourth Industrial
Revolution? In O.-A. Rognstad et al. (Eds.), Promoting Sustainable Innovation
and the Circular Economy: Legal and Economic Aspects (Forthcoming)
Routledge
OECD. (2018). Going Digital in a Multilateral World: An Interim Report to
Ministers, Executive Summary Meeting of the Council at Ministerial Level,
30–31 May 2018. Retrieved December 12, 2023, from http://www.oecd.
org/mcm/documents/C-MIN-2018-6-EN.pdf?utm_source=Adestra&utm_
medium=email&utm_content=Read%20the%20report%20here...&utm_
campaign=OECD%20Going%20Digital%20News%20-June%20
2018&utm_term=demo
Phillips A.A. (2006). Strengthen Pharmaceutical Patent Rights: Lowering the
Cost of Prescription Drugs by Stopping the Reckless Patent Litigation Abuse
of Generic Companies. 13 Connecticut Insurance Law Journal, 397–426
Podszun, R. and Pfeifer, C. (2022). Datenzugang nach dem EU Data Act: Der
Entwurf der Europäischen Kommission. Gewerblicher Rechtsschutz und
Urheberrecht, 953–961
Regulation (EU) 2016/679 of the European Parliament and of the Council of
27 April 2016 on the protection of natural persons with regard to the pro-
cessing of personal data and on the free movement of such data, and repeal-
ing Directive 95/46/EC (General Data Protection Regulation). Retrieved
December 12, 2023, from https://eur-lex.europa.eu/legal-content/EN/TXT/
PDF/?uri=CELEX:32016R0679
Reichman, J.H. (2009). Comment: Compulsory Licensing of Patented
Pharmaceutical Inventions: Evaluating the Options. 37 Journal of Law,
Medicine & Ethics, 247–263
Richards, D.G. (2008). A Skeptik’s View of Intellectual Property Rights. In
R.C. Bird & S.C. Jain (eds), The Global Challenge of Intellectual Property
Rights (pp. 267–282) Edward Elgar Publishing
Röttgen, C. (2018). Like or Dislike – Web Tracking. In T. Hoeren, & B. Kolany-
Raiser (Eds.), Big Data in Context: Legal, Social and Technological Insights
(pp. 73–79) Springer Nature
Seuba, X. (2021). Big Data, AI and Border Enforcement of Intellectual Property
Rights. In M. Burri (Ed.), Big Data and Global Trade Law (pp. 177–192)
Cambridge University Press.
Smith, R.D., Correa, C. and Oh, C. (2009). Trade, TRIPS, and Pharmaceuticals.
373 The Lancet, 684-691
Specht-Riemenschneider, L. (2022a). Der Entwurf des Data Act: Ein großer
Wurf in die falsche Richtung? GRUR Gewerblicher Rechtsschutz und
Urheberrecht, 937–938
Specht-Riemenschneider, L. (2022b). Data Act – Auf dem (Holz-)Weg zu mehr
Dateninnovation? Zeitschrift für Rechtspolitik, 137–140
Specht-Riemenschneider, L. (2022c). Der Entwurf des Data Act: Eine Analyse
der vorgesehenen Datenzugangsansprüche im Verhältnis B2B, B2C und
B2G. MMR Zeitschrift für IT-Recht und Recht der Digitalisierung, 809–826
Storr, C. & Storr, P. (2017). Internet of Things: Rights to Data from a European
Perspective. In M. Corrales et al. (Eds.), New Technology, Big Data and the
Law (pp. 65–96) Springer Nature
Thaler v Commissioner of Patents [2021] FCA 879
The European Commission. (2022, February 23). Proposal for a Regulation of the
European Parliament and of the Council on harmonised rules on fair access to
and use of data (Data Act). COM(2022) 68 final. Retrieved December 12,
2023, from https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CE

LEX:52022PC0068
The European Commission. (2022, February 23). Commission Staff Working
Document, Impact Assessment Report accompanying the document Proposal for a
Regulation of the European Parliament and of the Council on harmonised rules
on fair access to and use of data (Data Act). SWD(2022) 34 final. Retrieved
December 12, 2023, from https://digital-strategy.ec.europa.eu/en/library/
impact-assessment-report-and-support-studies-accompanying-pro-
posal-data-act
The European Commission. (2022, May 5). Proposal for a Regulation of the
European Parliament and of the Council on the European Health Data Space.
COM(2022) 197 final. Retrieved December 12, 2023, from https://eur-lex.
europa.eu/resource.html?uri=cellar:dbfd8974-cb79-11ec-b6f4-
01aa75ed71a1.0001.02/DOC_1&format=PDF
The European Parliament and the Council. (2023, November 15). Regulation
of the European Parliament and of the Council on harmonised rules on fair
access to and use of data and amending Regulation (EU) 2017/2394 and
Directive (EU) 2020/1828 (Data Act). PE-CONS 49/23. Retrieved
December 12, 2023, from https://eur-lex.europa.eu/legal-content/EN/TXT/
PDF/?uri=CONSIL:PE_49_2023_INIT
United Nations Economic and Social Council (ECOSOC). (2001, June 27).
The Impact of the Agreement on Trade-Related Aspects of Intellectual
Property Rights on Human Rights. Un Doc E/CN.4/Sub.2/2001/13
United Nations Secretary-General’s High-Level Panel on Digital Cooperation.
(2019). Report of the UN Secretary-General’s High-Level Panel on Digital
Cooperation: The Age of Digital Interdependence. New York/Geneva: United
Nations Publications
Thouvenin, F. and Tamò-Larrieux, A. (2021). Data Ownership and Data Access
Rights: Meaningful Tools for Promoting the European Digital Single Market?
In M. Burri (Ed.), Big Data and Global Trade Law (pp. 316–339) Cambridge
University Press
Veil, W. (2022). Ende des Wilden Westens oder wirrer Wildwuchs? ZGI
Zeitschrift für das gesamte Informationsrecht, 197–198
von Schönfeld, M. and Wehkamp, N. (2018). Big Data and Smart Grid. In
T. Hoeren, & B. Kolany-Raiser (Eds.), Big Data in Context: Legal, Social and
Technological Insights (pp. 93–106) Springer Nature
Legal Nature of NFTed Artwork:
A Comparative Study
Jia Wang and Arianna Alpini
1 Introduction
The platform economy is the tendency for commerce to move increas-
ingly towards digital platform business models. Platforms are underlying
computer systems that can host services allowing consumers, entrepre-
neurs, businesses and the general public to connect, share resources or sell
products. Blockchain technology can support immutable and trustless
transactions in a distributed and disintermediated way among various
users. The trustless environments that blockchains have created enable
peer-to-peer (P2P) sending and receiving transactions, smart contract
agreements, and more. On the blockchain platform, tokenisation is used
to transform ownerships and rights of particular assets into a digital form.
J. Wang (*)
Durham Law School, Durham University, Durham, UK
e-mail: jia.wang2@durham.ac.uk
A. Alpini
Department of Law, University of Macerata, Macerata, Italy
e-mail: arianna.alpini@unimc.it

https://doi.org/10.1007/978-3-031-42576-9_3
66 J. Wang and A. Alpini
Tokenisation in blockchain opens up multiple new possibilities for

businesses and individuals. Non-fungible tokens (NFTs) are widely
adopted by the token owner in the form of a record and hash codes that
show ownership of the unique token associated with a particular digital
asset. Transactions are executed on smart contracts, sequences of com-
puter codes that automatically execute pre-established instructions. Each
block has a set amount of storage capacity, and once it is filled, it is
‘chained’ to the previously filled block. Most NFTs exist on the Ethereum
blockchain, with permanent digital records of all cryptocurrency
transactions.
Blockchain technologies and platforms inspired the creation of ‘crypto
art’, that is, to tokenise artwork digitally and trade the tokens on the
platform. Crypto art is disruptive to the traditional market of artwork in
many ways. First, it solves the chronic problem of the piracy of artworks.
NFTs enable authors to have exclusive control of their works stored in
unique blocks with a private key, thus re-creating ‘digital scarcity’ of the
works. The transaction of NFTed artwork occurs within a more decen-
tralised power structure (Storey 2022). Second, NFTs make fine art
investment more accessible and democratic. On the one hand, it enables
a smaller investment of a fraction of an artwork. On the other hand, art-
work can also be directly accessible from mobile phones and laptops
without requiring storage space or the use of special equipment, which
thus allows trade to be conducted more seamlessly (Hashtag Investing
2022, Hamilton 2022). Third, it enables direct transactions between the
rightsholder and consumer without involving middlemen such as cura-
tors, galleries and art dealers (Drobitko 2022). Lastly, transactions of
NFTed artwork increase liquidity and allow for higher transparency of
data. Blockchain transactions are often completed in milliseconds, reduc-
ing the waiting time when selling NFTed artwork and allowing artists to
be paid more quickly, thus increasing liquidity (Storey 2022). Not only
are NFTs generally sold and traded in full public view (Storey 2022), but
each transaction can also be traced and followed, as each NFT and its
blockchain entries contain proof of current and past ownership, and all
transactions involved (Storey 2022).
While NFTs look promising in many respects in trading artworks on
digital platforms, legal risks must be addressed to capture the benefits of
NFTs fully. When buying an art piece, one does not purchase its copy-
right, which would have to be transferred separately. According to inter-
national copyright law, to own a piece of artwork does not necessarily
entail the subsequent right to display the work in a public place and col-
lect copyright royalties paid for the use of the work. The right to display
and receive royalties remains with the copyright holder.
In this chapter, we explore the legal nature of NFTed artwork. When
purchasing an NFTed artwork, what rights and interests does the pur-
chaser acquire? Is holding an NFTed work equivalent to holding a physi-
cal copy? The objectives of this chapter are, first, to contextualise the
discussion against the background of the trade of artworks in a block-
chain environment; second, to examine the most recent legislation and
court decisions concerning NFTs and cryptocurrency; finally, to conduct
a comparative study between the selected common law and civil law sys-
tems of their concept of assets and property in relation to NFTs.
We undertake a comparative approach to facilitate the understanding
of how different jurisdictions view the legal nature of NFTs. We look at
the Anglo-American countries, then turn to Europe and China as civil
law jurisdictions. Traditionally, the two legal systems view property rights
differently. A comparative study offers insight into whether the conver-
gence of law applies to legal issues of global relevance in a digital environ-
ment. Furthermore, this chapter conducts a case study of the most recent
court decisions concerning property and intellectual property issues. As
the legislation could lag behind the speedy advancement of technologies,
court decisions offer a more timely reflection of the judiciary’s attitude
to NFTs.
This chapter proceeds as follows. The first section introduces block-
chain and NFTs as an application of blockchain technologies, particu-
larly the tokenisation of artwork. The second section examines how
artworks are traded in the mortar-and-brick era and the digital era. The
third and fourth sections scrutinise the common law and civil law
approaches towards NFTs and the NFTed artwork. The last section offers
a comparative analysis of the different approaches and provides conclud-
ing remarks.
2 Blockchain and the Tokenisation

of Artwork
2.1 The Blockchains: Private, Public
and Consortium Chain
Blockchains have been developed into three types: private chain, public
chain and consortium chain. Private chains operate under the control of
certain individuals or organisations. This is done by setting up a permis-
sioned network, restricting the individuals allowed to participate in the
network and transactions (Jayachandran 2017). As participants must
obtain an invitation to join the network, private chains can filter out
illegal activities within a chain (Iredale 2021). Yet as there are fewer nodes
within the chain, the entire chain may become more easily compromised,
posing a potential security risk (Brown 2022). Having a unique hash is
vital for maintaining security throughout the different blocks because
there are concerns about hackers tampering with the blocks and changing
the hashes.
In contrast to a private chain, public chains, such as Bitcoin and
Ethereum, are openly accessible to all who can access the internet
(Jayachandran 2017), who can also see the ledger and participate in the
consensus process (Iredale 2021). This design enables participants in a
public chain to have equal rights and boosts transparency (Iredale 2021).
Yet it also comes with drawbacks. For instance, a high amount of compu-
tational power is necessary for a public chain to function, to maintain the
distributed ledger at a large scale (Jayachandran 2017). It takes a rela-
tively long time to verify each transaction compared to other chains.
A consortium chain is a permissioned ledger where information can
only be shared among a small group of organisations (Crypto News 2021).
It is formed by combining various private blockchains belonging to differ-
ent groups, where each group forms a node on the chain as a stakeholder.
While each group manages its own blockchain, the data within it can be
accessed, shared and distributed among the other organisations in the
consortium (Bybit Learn 2022). A consortium chain is created among
organisations to facilitate cooperation among these groups (Banerjee
2022). As a result, consortium chains offer benefits such as efficiency in

decision-making (Bybit Learn 2022). However, there are also challenges.
For example, a unified framework of industry standards for consortium
blockchains (Banerjee 2022) urgently needs to be developed.
2.2 The Tokens: Fungible and Non-fungible
An NFT is a piece of digital artefact that represents the ownership of real-

world assets. The influential NFT marketplaces in 2022 are OpenSea,
Rarible, NBA Top Shot, Binance and Nifty Gateway, with most of them
utilising cryptocurrencies as their payment method (Rodeck 2022).
Whilst NFTs operate as a type of cryptocurrency (Fairfield 2021), it is
vital to distinguish Bitcoin’s and NFT’s different natures. Bitcoins are
interchangeable and indistinguishable, making them fungible tokens
(Nakamoto 2019), while NFT is non-fungible because the associated
data has a unique “hash value”, a “unique and reproducible alphanumeric
value from a specific data set” (Tipotsch 2021) derived from the artwork.
There are different types of tokens. For example, using the ERC 20
Protocol, parties may create fungible tokens; using ERC 721 Protocol,
they may create non-fungible tokens; using ERC 1238 Protocol, they
may create non-transferable tokens (titles or badges).
There are three ways to create and issue tokens. The first method is
through an Initial Coin Offering (ICO). This is a method of raising capi-
tal for new ventures (Delivorias 2021). The tokens can be exchanged for
future products and services or confer a right to a share in future profits
on holders (Knowledge at Wharton 2019). The tokens are then launched,
and the business can use the proceeds to launch new products and ser-
vices (CFI 2022). The second method is mining, where groups or indi-
viduals compete to solve complex mathematical problems (Trading
Education 2021). The first one who solves the equation and validates the
accuracy of a transaction in a block wins a reward (Trading Education
2021). Upon mining, tokens can be minted, that is, published on the
blockchain and made available for purchase (Craig 2021). This can be
done on platforms such as OpenSea, which allows one to mint tokens on
the Ethereum blockchain by setting up a crypto wallet, creating a
collection and uploading work (OpenSea Learn 2022). The third method
to create tokens is through tokenisation by linking or embedding the
economic value and the rights derived from the asset to digital tokens
created on the blockchain (OECD 2020). The tokenised asset can then
be listed and sold on NFT marketplaces. In this chapter, we mainly focus
on the third tokenisation method for discussing NFTed artworks.
3 The Trading of Artworks

In the past, artworks were traded by transferring the physical object or
licensing the use of the work without the transfer of ownership. Collective
copyright licensing is managed by collective societies in Europe and
China while in common law jurisdictions, by corporations specialising in
collective copyright management. A common problem of collective
copyright management is the high agency cost. Moreover, tracing and
tracking the author of orphan works is less efficient due to the need for
more technology. The intermediary, that is, the collective society, is con-
fined by its structural and technological limits and cannot assist the right-
sholder in fully capturing the value of artworks.
In the digital era, various data management tools help improve the
efficiency of collective management. However, the agency cost and
source-tracing problems still exist, although to a lesser extent. With the
onslaught of blockchain and NFTs, it is time to revisit the agent-based
transaction model. Do artists need an intermediary to manage their
copyright if they can manage their own copyrighted works with secured
NFTs and supporting platforms?
Although blockchain technologies seem to offer an ideally distributed
and democratic digital world without intermediaries involving transac-
tion costs (Drobitko 2022), in the new reality of business, NFTed art-
work trading platforms emerge as new intermediaries. In China, the
platforms can be divided into three categories. The first type comprises
platforms that offer blockchain technology. Examples in China include
Zhixin Chain, a platform based on Hyperledger technology, which is
employed to execute smart contracts, digitise company chops and delin-
eate ownership of intellectual property. Beyond China, Hyperledger
Fabric, hosted by the Linux Foundation, offers enterprise-grade block-

chain technology for leaders in finance, banking and supply chains.
The second type is the platforms that support the trade of NFTed
work. These may include platforms that facilitate the collection and resale
of NFTed work, such as Netease Chain, a public chain which allows trade
in a large variety of NFTed work, such as animation, music and 3-D
models. Similarly, some provide the space for users to sell their NFTed
products, for instance, the Two-Mirror Museum (双镜博物), a public
chain that sells NFTed products related to culture, such as a digital col-
lectible featuring the Palace Museum’s only existing set of wedding gar-
ments belonging to a Qing dynasty empress. Such platforms can also be
found outside China, for instance, Nifty Gateway, which cooperates with
musicians and artists to create limited edition NFTs that are exclusively
available on their platform.
The third type relates to those platforms that issue NFTed work. An
example of such a platform is BlueFocus, a multinational company spe-
cialising in marketing and brand management services in media and
design. Another example is CrowdCreate, whose services include assis-
tance in crypto community management and crypto token marketing.
The increasingly involved artificial intelligence in creating artwork
brings another dimension to the legal compliance for the trading of art-
works. Unlike traditional artworks like painting, photography and music,
which are protected by copyright, AI-created art is less concerned with
copyright. AI-created art is also called generative art. Many NFT proj-
ects, such as CryptoPunks, Bored Ape Yacht Club, World of Women,
Azuki, Chromie Squiggles, Clone X, and Moonbirds, involve generative
art in the creation process. Generative art is generated wholly or in part
by the algorithm and not in direct control of the programmer, who is an
artist or commissioned by a customer. The programmer creates a pro-
gramme consisting of one or more algorithms that randomly generate an
artwork based on randomised parameter selections or by one or more
inputs that are driven or operated to suggest a direction for the artwork.
In more complex projects, artificial intelligence is programmed to make
decisions during the entire process of creating an artwork (Dornis 2021).
Different from the purchase of traditional artworks, the purchase of gen-
erative artwork is obtained upon the creation of the art, which coincides
with the act of minting. The purchase of the NFTed AI-created artwork
takes place by creating and minting the artwork.
However, NFTs are not the artwork, nor does it become the artwork.
The NFT records the existence and ownership of the artwork onto the
blockchain, and because no two NFTs are the same, and no two block-
chain registrations can be the same, the tokenised asset linked to the NFT
also can be considered unique and non-fungible. Each NFT contains
metadata that describes the corresponding assets in order to prove the
physical object’s authenticity or rarity. The NFT represents the physical
object in code written into the blockchain containing various informa-
tion. This information frequently contains the name of the creator of the
NFT, a URL linking to a representation of the underlying work of the
NFT, the date it was minted, and any contractual terms that follow the
NFT after it is sold. While the separate URL embedded in the NFT con-
tains a link to a copy of the underlying work, it is not itself a copy of that
work. Thus, an NFT is not a reproduction of content; it is merely a token
that authenticates the source of the content. For this reason, NFTs them-
selves are not “copies” and thus not subject to copyright infringement.
The metadata does not contain any recognisable content of the underly-
ing work, nor does it describe its contents. Similarly, the metadata does
not add, transform, or recast any underlying work.
Nevertheless, NFT establishes an exclusive ownership relationship
with the underlying artwork. The hash is stored on a blockchain with an
associated time stamp. Consequently, NFT keeps track of hash sales, so it
is possible to trace the hand steps of the hash to the creator. This mecha-
nism provides proof of authenticity and, simultaneously, ownership of
the work. The transfer of an NFT connected to a work of art transfers the
digital ownership of the authentic copy of the work; however, the pur-
poses that can be pursued with this tool are different, so it is necessary to
identify the crypto activity, the specific utility that the NFT is intended
to create from time to time.
Authors can create and sell NFTs representing their works. In practice,
the NFT digital art market recognises the owner of a “legitimate” NFTed
work as the “owner” of the work, even though NFTs do not convey copy-
right ownership of the work (Frye 2022). NFT owners encourage others
to use their work because popularity increases the value of the work.
Increasing the author’s impact creates more value than controlling the use
of the work. If the profit from selling NFTs alone is large enough to moti-
vate authors, copyright is no longer necessary as a legal monopoly to
reward authors. The value of art has always come from the reputation of
the author and the scarcity of the work through “authenticity”.
4 Legal Status of NFT: The Common

Law Approaches
The premise of this property syllogism is that “a particular type of right
(such as a chose in action, an intellectual property right or a beneficial
interest under a trust) is the same type of right as a right to a tangible asset
and must therefore be protected in the same way.” (McFarlane and Douglas
2022, page 162). With this background in mind, it is perhaps unsurprising
that crypto assets have been more readily accommodated within common
law systems’ vague notions of property than those of civilian systems.
The courts in England and Wales (English and Welsh cases), Singapore
(Singaporean cases), and New Zealand (Ruscoe v Cryptopia Ltd. (in Liq)
[2020]) have acknowledged bitcoins and other crypto assets as property
within the common law. Property in the case law to date refers more to
assets than things (Low and Hara 2022). This was also the advice of the
LawTech Delivery Panel’s UK Jurisdiction Taskforce in its Legal Statement
on Cryptoassets and Smart Contracts. The key question is how crypto
assets as property would fit within the common law’s classificatory scheme
for property. Unlike the civilian classification between movables and
immovables, the common law classifies property into real and personal,
with the former comprising mostly land. Personality is, in turn, classified
into either choses in possession or choses in action.
4.1 The United Kingdom
The point on whether cryptocurrency could be a form of property was

more fully developed in AA v Persons Unknown [2019] (“AA”). Bryan J
noted that the immediate difficulty was that “English law traditionally
views property as being of only two kinds, choses in possession and cho-
ses in action” (AA at [55], citing Colonial Bank v Whinney [1885]
(“Colonial Bank”)). Bitcoins, and other cryptocurrencies, did not fall
neatly into either category and thus could not be classified as a form of
property (AA at [56] and [58]). Bryan J, however, considered that it was
“fallacious to proceed on the basis that the English law of property recog-
nises no forms of property other than choses in possession and choses in
action” (AA at [58]). In doing so, he cited extensively from the legal state-
ment on crypto assets and smart contracts published by the UK
Jurisdiction Task Force (the “Legal Statement”). Thus, the Task Force
believed that Colonial Bank was not to be treated as limiting the scope of
what kinds of things could be property in law. Rather, it showed the abil-
ity of the common law to stretch “traditional definitions and concepts to
adapt to new business practices” (Legal Statement at [77]). The Legal
Statement, therefore, formed the basis for Bryan J’s conclusion that while
a crypto asset might not be a thing in action based on a narrow definition
of that term, it could still be considered property (AA at [59]). He made
a finding that crypto assets such as Bitcoin were property, given that they
met the four criteria set out in National Provincial Bank Ltd. v Ainsworth
[1965] (“Ainsworth”) at 1248 – namely that it must be “definable, iden-
tifiable by third parties, capable in its nature of assumption by third par-
ties, and have some degree of permanence or stability”.
The UK’s High Court recently ruled that NFTs are property, and thus
victims of NFT theft can now have their stolen assets frozen through
court injunctions. The decision comes after months of repeated NFT
thefts, as savvy hackers have exploited loopholes and poor security liter-
acy to seize high-profile NFTs. In an earlier case involving NFTs
(Osbourne v Persons Unknown [2022]), the court also found a claimant
has a good arguable case that misappropriated crypto assets are held on a
constructive trust is therefore clear that the courts are open to construc-
tive trust claims as regards crypto assets. However, Ms. Osbourne did not
go so far as to seek, as Mr. D’Aloia has, to ask the courts to consider a
claim in which – in addition to the alleged fraudsters - the exchanges are
also said to hold the crypto assets on constructive trust. In D’Aloia case,
therefore, appears to be the first in which the Court has found that there
is a good arguable case for this claim against the exchanges themselves.
D’Aloia case allows victims to obtain court injunctions against individu-

als whose crypto wallet has been identified as carrying a stolen NFT and
to the NFT platform on which the stolen asset is being sold.
This ruling demonstrates that the English courts are open to entertain-
ing constructive trust claims concerning crypto assets, not only against
the fraudsters themselves but also against third-party exchanges. The pos-
sibility of such a claim has been lent further support by the Law
Commission’s analysis in their Consultation Paper on Digital Assets,
published on 28 July 2022 (see paragraph 19.51). This would give vic-
tims of crypto-asset fraud a means of direct action against exchanges for
breach of trust should they fail to comply with their duties as construc-
tive trustees, having been notified that they are in the possession of fraud-
ulently misappropriated crypto assets.
4.2 Singapore
In an earlier case of CLM v CLN [2022] SGHC 46 (“CLM”), Lee Seiu

Kin J dealt with the question of whether stolen cryptocurrency assets,
specifically Bitcoin and Ethereum, could be the subject of a proprietary
injunction. Having considered the cases and the analysis in Ruscoe v
Cryptopia Ltd. (in liq) [2020] (“Ruscoe”), the judge was of the view (at
[46]) that the claimant, in that case, was able to prove an arguable case
that the stolen cryptocurrency assets were capable of giving rise to propri-
etary rights, which could be protected via a proprietary injunction.
In Janesh s/o Rajkumar v Unknown Person [2022], the court noted
that although NFTs have been characterised as certificates of ownership
“powered by smart contracts and protected by blockchain technology”
(Aksoy and Üner 2021, page 1115), NFTs represent an “ownership of a
digital certificate of authenticity of commonly available digital art” (Low
and Hara 2022). Nevertheless, the court disagrees with the ‘NFT is cer-
tificate’ approach. The court points out that NFTs are not just mere
information but rather data encoded in a certain manner and securely
stored on the blockchain ledger. (Janesh s/o Rajkumar v Unknown Person
2022, paragraph 58) Rather, NFTs provide instructions to the computer
under a system whereby the “owner” of the NFT has exclusive control
over its transfer from his wallet to any other wallet.
Lee Seiu Kin J adopted the Ainsworth test and upheld the following
findings. First, an NFT with its unique metadata is definable (Janesh s/o
Rajkumar, [44]). NFTs are not just mere information, but rather, data
encoded in a certain manner and securely stored on the blockchain ledger
(Janesh s/o Rajkumar, [58]) ‘It provides instructions to the computer
under a system whereby the “owner” of the NFT has exclusive control
over its transfer from his wallet to any other wallet.’ (Janesh s/o Rajkumar,
[58]) Second, per the second requirement that the “asset must have an
owner being capable of being recognised as such by third parties” (CLM,
[45(b)], citing Ruscoe at [109]) the presumptive NFT owner would be
whoever controls the wallet which is linked to the NFT an NFT with its
private keys would be an asset, with an owner being capable of being
recognised as such by third parties. The third requirement is that “that the
right must be capable of assumption by third parties, which in turn
involves two aspects: that third parties must respect the rights of the
owner in that asset, and that the asset must be potentially desirable” The
‘nature of the blockchain technology gives the owner the exclusive ability
to transfer the NFT to another party, which underscores the “right” of
the owner.’ Lastly, an NFT has a relevant degree of permanence and sta-
bility as money in bank accounts which, nowadays, exist mainly in the
form of ledger entries and not cold hard cash.
In summary, Singapore’s first court decision on NFTs rejects the anal-
ogy between a title deed and a certificate of property. The court applies
the Ainsworth criteria and upholds NFTs as property.
5 Legal Status of NFT: The Civil

Law Approaches
5.1 The Theoretical and Legislative Discussion
Scholars pointed out that digital tokens may be considered “digital

assets”. The same approach is taken by the Gesetz über Token und
VT-Dienstleister of Liechtenstein, which regulates tokens as assets
(Vermögen, art. 4) (Teruel 2021). In fact, it has been stated that “all-
European-legal systems of the member states regard not only corporeal
thing as the objects of real rights but also incorporeal assets, such as pat-
rimonial rights” (Von Bar and Drobnig 2004). However, some legal sys-
tems take a narrower approach concerning the scope of the property,
which is limited to “corporeal things”, as is the case under both German
and Swiss law (e.g. §§90 Bürgerliches Gesetzbuch30 -BGB- or §641
Swiss Civil Code 31). This means that tokens could not be regarded as
the object of property in these legal systems, e.g. in Germany, tokens have
been defined as “eine faktische Vermögensposition”, meaning ‘a factual
situation’ (Lehman and Krysa 2019). Other EU legal systems have either
incorporated a broader definition of the concept of a “thing” to include
“patrimonial or valuable rights” (e.g. arts. 334.10 CC; §§292, 298 and
299 Allgemeines bürgerliches Gesetzbuch, ABGB35) or a broader defini-
tion of the concept of an “asset” (art. 3.1 Burgerlijk Wetboek -BW-),
which makes regulating tokens as an object of ownership more accessible.
For example, in Spain, the judgment of the Supreme Court 20/06/201937
denied the recognition of bitcoin as a legal tender (money) but consid-
ered it an “incorporeal asset”. In Italy, tokens have been regarded as “digi-
tal assets” under the provisions of art. 810 Italian Civil Code (“Sono beni
le cose che possono formare oggetto di diritti”); and art. 65 French Loi n.
486 categorises tokens as “incorporeal assets” (bien incorporel).
From an EU perspective, the Proposal for a Regulation of the European
Parliament and of the Council on Markets in Crypto-assets (MiCA) 24
September 2020 aims to enhance legal certainty to crypto-assets while
encouraging innovation and protecting consumers. However, this pro-
posal does not cover the legal nature, the legal effects and the admissibil-
ity of using asset-backed tokens to transfer property rights. Tokens issued
in blocks of fewer than 150 tokens are excluded from the Regulation. So,
the Regulation does not cover small issuances, which are typically the
case in the tokenisation of real-world assets. The European Union Court
of Justice has ruled that cryptocurrencies fall into legal goods exempt
from VAT (EUCG, sez. V, 22 October 2015, case − 214/2016).
China’s first Civil Code became effective in 2021 and includes provi-
sions peripherally relevant to virtual assets. In Book I General Provisions,
Article 114 provides that civil subjects enjoy property rights (rights in
rem), which are the exclusive rights to directly dominate a particular
thing, including ownership, usufructuary rights, and security interests.
Article 116 is a Numerus Clausus that limits the types and contents of
property rights exclusively by law. Article 127 provides that the Civil
Code shall recognise the existing legal provisions for virtual property.
5.2 European Cases
Italian Supreme Court indicated that an equivalent function approach

could be taken to allow the court to treat NFTs similarly to cryptocur-
rency with similar qualities. It held that the legal qualities of bitcoins
depend on the purpose of the usage of the currency and the utility they
produce. If a virtual currency is used for speculative purposes, it will be
considered a financial product (security token). The Court identifies the
requirements to qualify the securities offered (in the specific case LWF
Coin) as financial instruments. The test includes the purpose of the use
of capital, the expectation of return, and the risk directly linked to the use
of capital. (Court of Cassation, Penal Section, 30 November 2021, n.
44337 and 22 November 2022, n. 44378).
The Court of Rome, IP Chamber (Docket No. 32072/2022) enjoined
Blockeras s.r.l. from any production, marketing, promotion and offering
for sale, directly and/or indirectly, in any way and form, of the NFTs and
digital contents and ordered the defendant to withdraw from the market
and remove from every website the NFTs and the digital contents associ-
ated or products in general covered by the injunction (Rome, 19 July
2022). The dispute concerned trademark infringement and unfair com-
petition practice, consisting of the unauthorised use of words or figura-
tive marks through the production, marketing and online promotion of
digital playing cards with images that reproduced footballers’ NFTs. The
distinctive signs in question show the image of former player Christian
(Bobo) Vieri wearing the Juventus shirt and the team’s name. The Court
did not express an opinion about the legal nature of NFT. Still, it stated
that the circumstance that Bobo Vieri played for Juventus and that he
granted permission to the use of his image through the creation of cards
reproducing the player with the different shirts of the teams in which he
played does not, therefore, exclude the need to request authorisation for
using the registered trademarks owned by the teams whose shirts and
names are reproduced. The decision indicated that the court viewed the
infringement of IP by NFTs as equivalent to an infringement made
through physical reproduction. Therefore, we can deduce that the court
considers NFT as equivalent to (intellectual) property.
At the European level, the Court of Rome is the first to order an
injunction to the creation and marketing of NFTs infringing registered
trademarks. It also ordered the NFTs to be removed from the trading
website. The decision represents a reference point at a global level at a
time when strong attention is paid, by all operators in the sector, to the
legal aspects of this new digital tool. It goes from who defines NFT as “a
digital not interchangeable good, such as a photograph, a song or a video,
whose property has been authenticated and stored in a database called
blockchain and which can be collected, sold and exchanged on various
online platforms” (Trevisi et al. 2022) to those who consider them as
“unique digital certificates, registered in a blockchain, used as a means to
register the ownership of an object, as a digital artwork or a collectible
object” (EUIPO 2022). The decision represents an implicit accreditation
of the interpretation - already adopted by the main national and interna-
tional offices, including EUIPO, for which Class 9 is the one for regis-
tered trademarks used to distinguish certain types of “digital goods”.
The Italian Court reiterates the provision of art. 97 of the Copyright
Law, relating to the permitted uses of the right to the image of a person,
does not extend to the use of trademarks possibly represented in the same
image. The same consideration, however, also introduces the probably
most important concept of the decision, which confirms the fact that the
creation of NFTs - which are “goods intended for commercial sale” -
requires specific authorisation from the proprietor of the trade mark, of
which it, therefore, constitutes a separate infringement and distinct from
the infringement constituted by the use of the trademark in the digital
images associated with the NFT.
This confirms the preference for a legal definition of NFT that under-
takes a dichotomy between the certificate and the content (Janesh s/o
Rajkumar v Unknown Person 2022 SGHC 264). Above all, it explains
the ratio of the same precautionary order, which is not by chance kept
well distinguished between NFTs and the corresponding digital content,
inhibits the “production, marketing, promotion and offer for sale, direct
and/or indirect, in any way and form” of, on the one hand “of the NFT
(non-fungible token)” and, on the other hand, of any other “digital con-
tent or product generally bearing the photograph, even modified, and/or
the Juventus trademarks, as well as the use of said trademarks in any form
and manner”. This judgement echoes the US court decision involving
Maison Hermes against the artist Mason Rothschild and Nike in relevant
goods traded in StockX, a second-hand market.
This dichotomy of NFT/digital content cannot be ignored in the latter
case. The judge needs to decide whether the creation of an NFT generates
an intrinsic value rather than a mere digital certificate of ownership of the
associated property, which, hypothetically, the person who mints and
uploads the NFTed work is the legitimate owner of the work. We can
deduce that even if someone possesses a good legitimately, it can be
unlawful for the legitimate owner to produce NFTs of the good protected
by IP rights. Since the judgments are still at a first summary level, we can
only speculate on this issue.
The Commercial Law Court of Barcelona (Visual Entidad De Gestion
De Artistas Plasticos/ Punto Fa, S.L. [2022] AJM B 1900/2022 –
ECLI:ES: JMB: 2022:1900, Juzgado de lo Mercantil n° 09 de Barcelona)
delivered one of the first judgments dealing with the relationship between
intellectual property and NFT. The decision involves the fast fashion
brand Mango and the Spanish collective society for artists VEGAP
(Visual Entidad de Gestión de Artistas Plásticos). In March 2022, Mango
exhibited a series of artworks created by Farkas, an artist, in a virtual
museum on the Decentraland, a Web 3.0 site. Mango legitimately owns
the original copy of the works. The collection was designed to reinterpret
rather than directly reproduce the artworks, which are under copyright
protection. VEGAP sued Mango for copyright infringement, arguing
that the minting and displaying of the artworks infringed copyright;
Mango argued that the NFTs were just a list on OpenSea and did not
represent any proprietary rights per se. The Court ordered tokens to be
de-listed, and further pointed out that the withdrawal of a work does not
amount to destroying tokens since tokens can be used during the process.
For this reason, the Court orders the claimant to provide a cryptocur-
rency wallet, with a deposit of EUR 1000 that will be used to maintain
legal custody of the NFTs and orders OpenSea to transfer custody of

those tokens to be deposited to the applicant’s portfolio.
The first aspect of the dispute is the extent of Mango’s rights as owner
of the physical artwork. The ruling states that VEGAP transferred the
right to display physical works publicly but nothing else. The Court
assumes that the right to display does not give the right to digitise the
work, display and sell it as NFT. The second question that the Court will
have to examine is whether adapting a work in this way infringes copy-
right. Styles and ideas are not protected; only the expression of the idea is
protected. If the Court decides that these designs are indeed in violation
of relevant IP rights, the question is whether the minting NFT of a work
without authorisation is unlawful in itself and whether the display and
sale of such an NFT is a communication o the work to the public. It can
be argued that an NFT may include a link to a copy of the work but not
the work itself. The connection could not even be permanent and may be
interrupted. In addition, the actual connection to work is not always easy
to reach. If the link is to an IPFS file, it is not accessible unless a special-
ised browser like Brave, which can read IPFS links, is used. From this
perspective, it is difficult to admit that “the act of minting” is protected
as an exclusive right of the author.
5.3 Chinese Cases
In the first NFT court decision in China (Qice Technology Ltd. v A

Technology Ltd), the Hangzhou Internet Court, an intermediate-level
court, holds an NFT-trading platform liable for copyright infringement
for an NFTed unauthorised reproduction of an artwork uploaded by its
user. The court discusses similar issues regarding the nature of NFTs. It
holds that the metadata exclusively and uniquely represents an artwork
copy. It is identifiable by third parties and maintains the scarcity of the
work in a digital form. Therefore, an NFTed work is a ‘digital commod-
ity’, and NFTed copies of the work are digital assets. The trade of an
NFTed work is essentially a transfer of ownership of the copy being
tokenised and uploaded (page 18). Acquiring such an NFTed work
entails obtaining the property rights and interests in that copy. It entails
no license to use such digital assets nor a license or a transfer of the intel-
lectual property rights of the underlying artwork (unless the sales agree-
ment provides otherwise). In the further analysis of copyright
infringement, the court distinguishes NFTed work from a physical object.
It holds that the distribution right does not apply in this case because it
only concerns the distribution of physical objects. The legitimate creator
of an NFT should not be the person who possesses a copy of the underly-
ing work but the person who owns the copyright or obtains a due license
for the underlying work. Hence, it holds that the uploading of the NFTed
work infringed on the right to disseminate work by information networks.
However, as the court is only a district-level court, it remains to be seen
whether its ruling will be widely followed or is likely to be challenged in
subsequent cases by other courts in China. In any case, as the authorities
have not yet enacted any formal NFT laws or regulations, the court’s
insights in the judgment are meaningful, and NFT players in China
should carefully consider the implications of the ruling.
6 Conclusions
The value of art has always come from the reputation of the author and
the scarcity of the work through ‘authenticity’. NFTs offer artists the
opportunity to secure incomes with tracing and tracking functions and
embedded smart contracts while encouraging the dissemination of art-
work that cannot be reproduced without authorisation. If the profit from
selling NFTs is large enough to motivate authors, copyright as a legal
monopoly is no longer necessary to generate rewards. Recognising NFTs
as property encourages artists to be open to the market, which helps cre-
ate cultural prosperity and increase social welfare.
The NFT is more than a recording of digital work. Minting an NFTed
work is to record the work on the blockchain through an identification
code. The creation of the digital tokenised work (i.e. registered block-
chain with hash code) involves the acquisition of ownership by the regis-
tration holder. NFTed artwork might evoke the question of exclusive
possession and control of the work, including property and IP rights. The
right to tokenise a work protected by copyright belongs to the owner of
the copyright or those who have the authorisation of the owner; beyond
this hypothesis, this right belongs to the owner of NFTed work.
Consequently, it is necessary to distinguish NFTs from NFTed artwork.
The NFT is a mechanism for forming the ownership of a right of use and
disposition of digital work in the hands of the person who registers
the NFTs.
Although the judiciary from different jurisdictions has been willing to
extend the protection for brick-and-mortar property to NFTed artworks,
NFTs are at the risk of misrepresenting or even infringing IP rights in a
work minted into NFTs. Possessing an NFT does not necessarily confer
any legal right over the digital or physical object the NFT refers to. Several
proposals have been advanced to overcome this limitation to the concept
of NFT. Some try to strike a balance between the legal and the technical
dimension, incorporating aspects of copyright law into the metadata of
the NFT or in accompanying documentation; others propose to incorpo-
rate the actual work into the underlying smart contract. While many
commentators are critical at this point (Ryan 2021), others, like Fairfield,
see the potential of NFTs as forms of ‘unique digital property’, re-
establishing personal property rights that have been lost to user agree-
ments and other instruments of uneven bargaining power (Fairfield 2021).
In the UK, the Task Force took the view that Colonial Bank was not
to be treated as limiting the scope of what kinds of things could be prop-
erty in law. Instead, it showed the ability of the common law to stretch
“traditional definitions and concepts to adapt to new business practices”.
National Provincial Bank Ltd. v Ainsworth [1965] has established a four-
factored test for the property as something “definable, identifiable by
third parties, capable in its nature of assumption by third parties, and
have some degree of permanence or stability”. In Singapore, the court
rejected the analogy between a title deed for real property and NFTs as a
certificate for digital assets. Rather, it adopted the Ainsworth criteria and
upheld NFTs as property.
In Europe, although some jurisdictions traditionally take a narrow
approach concerning the scope of the property, which is limited to “cor-
poreal things”, many others are more open to including incorporeal assets,
such as patrimonial rights, into real rights. The court decisions discussed
in Sect. 5 demonstrate that cryptocurrency and NFTs are considered
digital assets (Calzolaio 2020). The Italian and Spanish Courts noted that
NFTs, on the one hand, bear a certifying function and, on the other hand,
link to specific content. China’s first Civil Code explicitly recognises vir-
tual property as an object protectable by law. In the first NFT-concerned
case, the court distinguishes the NFTed work from a physical object and
holds that an NFTed work is a ‘digital commodity’, and NFTed copies of
the work are digital assets. It adopts a test similar to the Ainsworth test
that evaluates whether the data is unique and securely linked to a work
and is identifiable by third parties. It is consensus that while the right of
the owner to mint and tokenise a work is to be fully protected, it cannot
be overlooked that this right must be exercised in compliance with the
principle of economic solidarity and fair competition.
The comparison between the legal systems implies an increasing level
of convergence of law towards a harmonised concept of digital assets in a
world built on blockchain and tokens. In comparing the common law
and civil law systems, convergence of law is emerging in the digital world.
The legal systems have advanced closer towards a concept of property
with many shared features. First, the data must be stored securely on the
blockchain ledger. The nature of blockchain technology gives the owner
the exclusive ability to transfer the NFT to another party, which under-
scores the “right” of the owner. Second, the data should be capable of
being recognised by third parties. Third, the data has intrinsic value that
is respected and potentially desirable by third parties. Lastly, an NFT has
a relevant degree of permanence and stability as money in bank accounts
which, nowadays, exist mainly in the form of ledger entries and not cold
hard cash. Property is moving from a static concept to a concept of act
and activity. At the same time, ownership is the link between the owner
and the worthy interest to be realised and guaranteed.
NFTed artworks are considered incorporeal assets that confer quasi-
property rights and interests. The legal principle of numerus clausus is an
instrument for legal certainty. However, reality goes beyond the dogmas
of the legal tradition. The jurist must take an evolutionary leap forward
to adapt the mechanisms and techniques of law to emerging digital tech-
nologies that change the societal ecosystem. The concept of property per
se and the rights deriving from the property need a recalibration that
shifts from focusing on exclusive control to the use of the thing.
Attention must be paid to the consideration that the “thing” becomes

a juridical good as a reference point and content of legal situations if it
has a socially appreciable utility and finds in the orderly system an evalu-
ation in terms of merit (Perlingieri 1990). The artificial exclusivity – cre-
ated by the relationship between the account/portfolio and the hash via
NFT – that represents the copy of the underlying value – evokes the
proprietary logic even if its dynamics undermine any theoretical defini-
tion for the benefit of the valuation of the interests involved.
References
Articles
Aksoy, P. C. & Üner, Z. (2021). NFTs and Copyright: Challenges and

Opportunities. Journal of Intellectual Property Law & Practice, 16(10),
1115–1126.
Bybit Learn (2022). ‘Consortium blockchain’, available at https://learn.bybit.com/
glossary/definition-consortiumblockchain/, accessed on 20 September 2023.
Calzolaio, E. (2020). Il bitcoin come oggetto di property. Note a margine di una
recente sentenza della High Court. Il Foro Italiano, CXLV 10, 494–500.
De Caria, R. (2020). Blockchain and Smart Contracts. Italian Law Journal,
6(1), 363–380.
De Caria, R. (2018). The Legal Meaning of Smart Contracts. European Review
of Private Law, 26(6), 731–752.
Dornis, T.W. (2021). Of ‘Authorless Works’ and ‘Inventions without Inventor’ –
The Muddy Waters of ‘AI Autonomy’ in Intellectual Property Doctrine.
European Intellectual Property Review, forthcoming 2021, 1–28
Fairfield, J. (2021). Tokenized: The Law of Non-Fungible Tokens and Unique
Digital Property. Indiana Law Journal, forthcoming. https://ssrn.com/
abstract=3821102
Frye, B. L. (2022). After Copyright: Pwning NFTs in a Clout Economy. The
Colombia Journal of Law and Arts, 45(3), 341–353.
Lehman, M., & Krysa F. (2019). Blockchain, smart contracts und token aus der
sicht des (internationale) privatrechts. Bonner Rechts Journal 2019 (Issue
2): 90–96.
Moringiello, J. M. & Odinet, C.K. (2022) The Property Law of Tokens. Florida
Law Review, 74, 607–671. https://ssrn.com/abstract=3928901 or https://
doi.org/10.2139/ssrn.3928901
Murray, M.D. (2022a). Generative and AI Authored Artworks and Copyright
Law. Hastings Communications and Entertainment Law Journal, 45(1),
28–43. https://ssrn.com/abstract=4152484 or https://doi.org/10.2139/
ssrn.4152484
Murray, M.D. (2022b). NFT Ownership and Copyrights. https://ssrn.com/
abstract=4152468 or https://doi.org/10.2139/ssrn.4152468
Nakamoto, S. (2019). Bitcoin: A Peer-to-Peer Electronic Cash System, ‘Bitcoin:
A Peer-to-Peer Electronic Cash System’, https://papers.ssrn.com/sol3/papers.
cfm?abstract_id=3440802, accessed on 20 September 2023.
Perlingieri, P. (1990). L’informazione come bene giuridico. Rass. dir. Civ.,
326–338.
Rodeck & Hooson. (2022). ‘Top NFT Marketplaces of 2023’, available at
https://www.forbes.com/uk/advisor/investing/cryptocurrency/best-nft-mar-
ketplaces/, accessed on 20 September 2023.
Ryan, T. R. (2021, December 2). Will the Artworld’s NFT Wars End in Utopia
or Dystopia? Art Review. https://artreview.com/will-the-artworld-nft-wars-
end-in-utopia-or-dystopia/.
Teruel, G.M.R., Simon-Moreno, H. (2021). The digital tokenization of prop-
erty rights: a comparative perspective. Computer Law and Security Review,
41, 4–5.
Von Bar C., & Drobnig, U. (2004). ‘The Interaction of Contract Law and Tort
and Property Law in Europe: A Comparative Study’, Sellier de Gruyter
(2004). Study on property law and non-contractual law as they relate to
contract law, European Commission–Health and Consumer Protection
Directorate-General, Brussels, Belgium, available at: https://www.academia.
edu/1131084/Study_on_property_law_and_noncontractual_law_as_they_
relate_to_contract_law (last access: 7 November 2016).
Book Chapters
Low, K. F. K. & Hara, M. (2022). Cryptoassets and Property. In Sjef van Erp &
Katja Zimmermann (eds), Edward Elgar Research Handbook on EU Property
Law (forthcoming), available at SSRN: https://ssrn.com/abstract=4103870
or http://dx.doi.org/10.2139/ssrn.4103870
Cases
AA v Persons Unknown, [2019a] EWHC 3556 (Comm)

Colonial Bank v Whinney [1885] 30 Ch.D 261
CLM v CLN [2022] SGHC 46
Janesh s/o Rajkumar v Unknown Person [2022] SGHC 264
Juventus Football Club S.p.A. v Blockeras s.r.l. [2022a] Court of Rome, IP
Chamber, Docket No.32072
National Provincial Bank Ltd v Ainsworth [1965] AC 1175
Osbourne v Persons Unknown & Anor [2022] EWHC 1021 (Comm) (10
March 2022)
Qice Technology Ltd v A Technology Ltd ([2022] Zhe 0192 Min Chu No. 1008
Ruscoe v Cryptopia Ltd (in Liq) [2020] NZHC 728, [2020] 2 NZLR 809
Skatteverket v Hedqvist [2015a] EUCG, sez. V, 22 October 2015, case -
264/2014
Visual Entidad De Gestíon De Artistas Plásticos v Punto Fa, S.L. [2022a] AJM
B 1900/2022 - ECLI:ES: JMB: 2022:1900, Juzgado de lo Mercantil n° 09 de
Barcelona
English and Welsh Cases
AA v Persons Unknown [2019b] EWHC 3556 (Comm)

Danisz v Persons Unknown [2022] EWHC 280 (QB) at [13]
Fetch.AI Ltd v Persons Unknown Category A [2021] EWHC 2254 (Comm) at [9]
Ion Science Ltd v Persons Unknown (unreported EWHC(Ch) 21 December
2020, Butcher J) at [11]
Mr Dollar Bill Ltd v Persons Unknown [2021] EWHC 2718 (Ch) at [10]
Shair.Com Global Digital Services Ltd v Arnold 2018 BCSC 1512 at [15]
Vorotyntseva v Money-4 Ltd [2018] EWHC 2596 (Ch) at [13]
Wang v Darby [2021] EWHC 3054 (Comm) at [55]
Singaporean Cases
B2C2 Ltd v Quione Pte Ltd [2019] SGHC(I) 03 [2019] 4 SLR 17 at [142], but
the issue was left open on appeal: [2020] SGCA(I) 02, [2020] 2 SLR 20 at
[144]. See now, however, CLM v CLN [2022] SGHC 46 at [46].
European Cases
Juventus Football Club S.p.A. v Blockeras s.r.l. [2022b] Court of Rome, IP

Chamber, Docket No.32072
Skatteverket v Hedqvist [2015b] EUCG, sez. V, 22 October 2015, case -
264/2014
Visual Entidad De Gestíon De Artistas Plásticos v Punto Fa, S.L. [2022b] AJM
B 1900/2022 - ECLI:ES: JMB: 2022:1900, Juzgado de lo Mercantil n° 09 de
Barcelona
Others
Banerjee, A. (2022). Everything you need to know about consortium blockchain.

Retrieved 4 November, 2022, from https://www.blockchain-council.org/
blockchain/everything-you-need-to-know-about-consortium-blockchain/
Brown, E. (2022). Comparing public and private blockchain features, pros &
cons. Retrieved 5 November, 2022, from https://bsvblockchain.org/news/
comparing-public-and-private-blockchain-features-pros-cons
Bybit Learn (2022). Consortium blockchain. Retrieved 4 November, 2022,
from https://learn.bybit.com/glossary/definition-consortium-blockchain/
CFI (2022a). Initial Coin Offering (ICO). Retrieved 6 November, 2022, from
https://corporatefinanceinstitute.com/resources/cryptocurrency/
initial-coin-offering-ico/
CFI (2022b). Minting Crypto. Retrieved 6 November, 2022, from https://cor-
poratefinanceinstitute.com/resources/cryptocurrency/minting-crypto/
Craig, J. (2021). Crypto Minting vs Mining: What’s the Difference? Retrieved 6
November, 2022, from https://phemex.com/blogs/crypto-minting-vs-crypto-
mining
Crypto News (2021). What is Consortium Blockchain? Retrieved 4 November,
2002, from https://crypto.news/glossary/consortium-blockchain/
Delivorias, A. (2021). Understanding Initial Coin Offerings. European
Parliamentary Research Services. Retrieved 6 November, 2022, from
https://www.europarl.europa.eu/RegData/etudes/BRIE/2021/696167/
EPRS_BRI(2021)696169_EN.pdf
Drobitko, A.(2022). Can Artists Still Benefit From NFTs? Retrieved 6 November
2022, from https://www.forbes.com/sites/forbesbusinesscouncil/2022/06/29/
can-artists-still-benefit-from-nfts/?sh=57efba0a648d
EUIPO, EUIPO Draft Guidelines 2023 edition. Retrieved 13 January, 2023,

from https://euipo.europa.eu/ohimportal/nl/draft-guidelines 2023
Hamilton, A. (2022). The Private and Public Benefits of NFT Art. Retrieved 6
November 2022, from https://www.zenofineart.com/blogs/news/
the-private-and-public-benefits-of-nft-art
Hashtag Investing (2022). Top 6 Benefits Of NFT’s For Artists And Reasons To
Use Them In 2022. Retrieved 6 November, 2022, from https://www.hashtag-
investing.com/blog/top-benefits-of-nfts-for-artists
Iredale, G. (2021). Public vs Private Blockchain: How do they differ? Retrieved
4 November, 2022, from https://101blockchains.com/public-vs-private-
blockchain/
Jayachandran, P. (2017). The difference between public and private blockchain.
Retrieved 4 November, 2022, from https://www.ibm.com/blogs/block-
chain/2017/05/the-difference-between-public-and-private-blockchain/
Knowledge at Wharton (2019). How to Design an Effective Initial Coin
Offering. Retrieved 6 November, 2022, from https://knowledge.wharton.
upenn.edu/article/design-effective-initial-coin-offering/
McFarlane, B. and Douglas, S. (2022). Property, Analogy and Variety. Oxford
Journal of Legal Studies, 42, 161–186.
OECD (2020). The Tokenisation of Assets and Potential Implications for
Financial Markets. OECD Blockchain Policy Series. Retrieved 6 November,
2022, from https://www.oecd.org/finance/The-Tokenisation-of-Assets-and-
Potential-Implications-for-Financial-Markets.pdf
OpenSea Learn (2022b). What is minting? Retrieved 6 November, 2022, from
https://opensea.io/learn/what-is-minting-nft
Rarible (2022). How do I sell my NFT on Rarible? Retrieved 6 November,
2022, from https://rarible.com/how-it-works/using-rarible/how-do-i-sell-
my-nft-on-rarible
Reiff, N. (2022). Decentralised Autonomous Organisation (DAO): Definition,
Purpose and Example. Retrieved 5 November, 2022, from https://www.
investopedia.com/tech/what-dao/
Storey, A. (2022). 11 Surprising Benefits of NFTs for Artists. Retrieved 6 November,
2022, from https://postergrind.com/11-surprising-benefits-of-nfts-for-artists
Tipotsch, A., Formulating a Smart Contract and Minting an NFT. Retrieved 13
January 2023, from https://www.schoenherr.eu/content/formulating-a-smart-
contract-and-minting-an-nft/
Trading Education (2021). Cryptocurrency Mining Explained. Retrieved 6

November 2022, from https://trading-education.com/cryptocurrency-mining-
explained
Trevisi, C., Visconti, R.M., Cesaretti, A. (2022). Non-Fungible Tokens (NFT):
business models, legal aspects, and market valuation. Rivista di Diritto dei
Media. Retrieved 13 January, 2023, from https://www.medialaws.eu/rivista/
non-f ungible-t okens-n ft-b usiness-m odels-l egal-a spects-a nd-m arket-
valuation/
Whittaker, M. (2022). How Does Bitcoin Mining Work? Retrieved 6 November
2022, from https://www.forbes.com/advisor/investing/cryptocurrency/bitcoin-
mining/
Intellectual Property Regulation
of Artificial Intelligence: A Matter
of Time or a Step Too Far?
Lucius Klobucnik
1 Introduction
Artificial intelligence (AI) has been disrupting the world of technology
and computing for many years, bringing significant benefits to our daily
lives. Ever more sophisticated AI systems do not only supplement the
work of humans, but can perform tasks which have been traditionally
reserved exclusively for humans. AI can be subdivided into a weak (or
narrow) AI and strong AI, according to whether it is designed to tackle
single task or can accomplish tasks across multiple domains. AI has been
increasingly involved in creative and inventive processes which result in
assets qualifying for intellectual property protection, whether by copy-
right or patent law. However, intellectual property law was designed to
protect inventions and creations of a human mind, thus seemingly leav-
ing AI outside of the scope of protection. Currently, the prominence of
AI has been set to shake up the foundations of intellectual property law
L. Klobucnik (*)
Aston Law School, Aston University, Birmingham, West Midlands, UK
e-mail: l.klobucnik@aston.ac.uk

https://doi.org/10.1007/978-3-031-42576-9_4
92 L. Klobucnik
(IP) law. IP law was established in order to recognise the fruits of human
ingenuity, creativity, and inventiveness. Indeed, for a long time, mankind
assumed that only humans are capable of producing creative works or
inventions. But what happens if the human element is taken away from
the creative or inventive process? Can IP law accommodate the ingenuity
of machines or will legislative changes be required? One of the widely
discussed questions has been whether AI can generate inventions or
copyright- protected works without human contribution. Suggestions
have been provided by academics, governments, policy makers, interna-
tional institutions (most notably the World Intellectual Property
Organisation – WIPO), but legislators have been hesitant to reflect any
of them into the existing legal framework. Policy discussions have been
revolving around the question whether intellectual property law, espe-
cially copyright and patent law, ought to be revised to account for tech-
nological developments in AI. For instance, the European Parliament, in
its resolution from February 2017 noted that ever more sophisticated AI
requires the legislation in virtually every area to consider its legal and
ethical implications without stifling innovation. In the UK, both copy-
right and patent law systems lack specific regulation regarding eligibility
of robot generated works or robot inventions. The UK Government set
out a goal to change this, by asking the public in Autumn 2021 whether
the current rules on copyright and patent are sufficient to accommodate
creations and inventions by AI, or whether a new set of rules is needed. A
large variety of stakeholders exchanged their views with the UK
Government. This may lead to new policy options and potential changes
in law, which will be discussed here. This chapter does not aim to provide
responses as to whether copyright and patent law is ripe for amendment.
Rather, it aims to discuss involvement of AI in the IP value chain – not
only with respect to AI as a creator/inventor (i. e. before IP protection
arises), but also the role of AI as a user of IP protected assets. Challenges
brought by AI are discussed here with respect to two areas of intellectual
property – copyright and patents.
Intellectual Property Regulation of Artificial Intelligence… 93
2 AI and Copyright

2.1 Originality of Works Created by AI
Perhaps the biggest challenge with respect to copyright-protected works

created by AI lies in the question whether these works will be able to meet
the central criterion for copyright protection – originality. This criterion
may be hard to reconcile with AI-generated works. Especially the test of
originality, established in the InfoPaq case and applicable in the UK and
the EU, is based on the premise of an author’s own intellectual creation,
thus presupposing a human quality. The criterion of originality focuses
on the interaction between the author and the work and has an inher-
ently human element. The human authorship requirement raises ques-
tions with regard to AI creations. In particular, when robots start operating
more independently, making their choices less predictable and foresee-
able to either their users or their programmers. Human involvement in
the process of creation becomes more difficult to identify. McCutcheon
suggests that the criterion of originality would be applied on a hypotheti-
cal basis – if the work had been authored by a human, or if that human
could be identified, would it be original? The test of originality is deeply
rooted in copyright legislation and in case law. Copyright exceptions for
the sake of AI-created works are currently not being considered.
Pursuant to the seminal InfoPaq decision, which plays a leading role in
understanding the originality requirement, the Court of Justice of the
European Union (CJEU) pointed in the direction that a creative output
has to have a human input. Answering the question whether eleven words
could enjoy copyright protection, the CJEU clarified that it was ‘only
through the choice, sequence and combination of those words that the
author may express his creativity in an original manner and achieve a
result which is an intellectual creation’. In defining the elements of origi-
nality, the CJEU did not look into the result of creativity but rather
focused on the process of creation. The clear focus of the CJEU on the
human element was confirmed in the case of Painer, where the CJEU
explained that an intellectual creation was author’s own if it reflected the
author’s personality. The process of creation of a copyright-protected
94 L. Klobucnik
work requires the author to make free and creative choices, which ulti-
mately reflect the author’s personal touch. Following these criteria, it
seems difficult to detach a copyright protected work from a human ele-
ment. Although the CJEU does not expressly refer to a human touch as
an authorship requirement, it indicates quite clearly that some kind of a
human quality has to be present in the creation process. According to
Ramalho, where there is no human author, a work cannot be original,
and without originality, a work cannot be protected by copyright.
The concept of originality underlines the need for a human author,
insofar as personality can be described as a purely human attribute. With
respect to the requirement concerning free and creative choices, Van
Gompel argues that even though the autonomy of authors to make free
choices is naturally restricted, as individual creators operate within social,
technical and institutional environments, at least some of these choices
are internal and self-imposed. The underlying rationale for free and cre-
ative choices seems to be that these choices are made consciously, i. e. by
a conscious being. Moreover, if copyright protects expression of ideas, it
would be questionable to think of an idea as being expressed by an AI
machine.
The international legal framework for copyright, namely the Berne
Convention, does not explicitly define authorship, nor does it explicitly
require a human author. However, one of the prominent international
copyright commentators, Sam Ricketson, argues that the Berne
Convention is very ‘anthropocentric’, e. g. by granting moral rights to the
author, by linking the duration of protection to the life of the author or
by referring to ‘intellectual creations’. The logical conclusion seems to be
that attributing moral rights to an AI machine or to link copyright pro-
tection to the ‘life’ of a subject matter would not make sense. The view
that copyright-protected works can only have a human author is corrobo-
rated by the International Association for the Protection of Intellectual
Property (AIPPI), which argues that without human intervention, there
should be no copyright protection of AI-generated works. Indeed, with
its robust exclusive rights and a long term of protection, copyright may
not be the most suitable vehicle for the protection of AI-generated works.
However, the copyright law framework may offer a potential vehicle for
protection of AI-generated works in the form of neighbouring
rights (related rights) – with a shorter duration and without the element
of moral rights, as opposed to copyright. A legislative construct that may
serve as an example is already present in the UK copyright law – computer-
generated works. Let us have a closer look at the relevance of the debate
on computer-generated works for AI-generated works.
2.2 Potential Legislative Lead Given by

Computer-Generated Works
One of the main questions related to AI and copyright is who is the

author (and thus the owner) of AI-*generated works. In the United
Kingdom, the Copyright, Designs and Patents Act (CDPA) 1988 already
has a rather unique provision on the so-called computer-generated works
(CGWs), which was added to the UK copyright law as a response to
technological challenges, i. e. a (digitally) created work that seemingly has
no author, or at least no human author. In case of CGWs, the author for
the purposes of copyright law is the person by whom the arrangements
necessary for the creation of the work are undertaken. When dealing with
an ‘artificial author’, the UK opted for a legislative route where the natu-
ral person behind the computer rather than the actual maker (the com-
puter itself ) is to be considered an author for copyright law purposes. The
difference between CGWs and other copyright-protected works is that in
the case of the former, the human author does not benefit from moral
rights and the term of protection is shorter (50 years from the end of the
calendar year of the creation). Such a right resembles a neighbouring
right rather than copyright in the sense of an author’s right, although the
CDPA does not use such a language. The legislative construction of
CGWs leaves some questions on authorship open to interpretation. For
instance, who is the person by whom the arrangements necessary for the
creation of the work have to be undertaken – will it be the person over-
seeing the operation of the machine, or the person who provides the
actual device, or the computer programmer who wrote the underly-
ing code?
96 L. Klobucnik
3 An Appropriate Mechanism to Use

Copyright-Protected Works by AI
3.1 AI Creator Learning from Existing
Copyright-Protected Works
The copyright law framework seems to suggest that a human element is

indispensable in creation of copyright protected works. However, how do
AI-generated works get created in the first place? It has to be noted that
in order to come up with artistic outputs, AI systems first have to engage
with already existing works. Put simply, AI has to be fed data (or works)
to come up with a creation. Technically, AI is combining different works
together to create a new work. Such a process poses a question as to
whether the use of copyright-protected content for AI training purposes
can be considered copyright infringement. Some copyright owners have
already expressed their discontent with AI using their works for training
purposes. Some of the biggest online picture repositories – GitHub and
Getty Images – have opposite views on whether AI training amounts to
infringement. While the former calls for an exception from copyright
protection and suggests a contributor fund to remunerate authors, the
latter decided not to offer AI-generated works on their platform at all. AI
tools can generate works, which resemble particular art styles of concrete
artists as a result of those works being fed into AI machine learning.
These are the so-called ‘text-to-image’ AI systems – in the past, they have
been available only to a select group of tech insiders. Recently, an image
generated with an AI system called Midjourney won an art competition
at the Colorado State Fair. In order for a machine to generate an image,
it must be trained on large amount of metadata, which may include bil-
lions of images scraped from the internet, paired with written descrip-
tions. On the one hand, copyright owners should be remunerated if their
works are being used. On the other hand, it would be prohibitive to
licence every single image in a dataset before using it. If we opt for the
licensing way, machine learning may become impossible. The concept of
fair dealing, known in UK law, may potentially be used for AI-generated
content based on AI training. Recently, there have also been some
suggestions by the UK Government to introduce a data-mining excep-

tion from copyright infringement for the purpose of AI-training.
However, the legislator should carefully weigh for such an exception not
to go so far as to suggest that artworks fed into AI systems can be used
without the rights holder’s consent. Let us have a closer look at the pro-
posed exception and its potential legal implications.
3.2 Text and Data Mining (TDM) Exception
3.2.1 Text and Data Mining Exception Promoted by

the UK Government
The UK Government had been planning to introduce a new copyright

and database exception which allows text and data mining for any pur-
pose, commercial or non-commercial. It has to be noted, however, that
these plans have been halted due to a significant backlash from the cre-
ative industries. Therefore, the fate of the proposed text and data mining
exception remains unclear and the UK is currently left with the exception
only covering non-commercial research, which is generally more restric-
tive than in other jurisdictions around the world. In any case, the pro-
posed text and data mining exception deserves our attention. Text and
data mining (TDM) means using computational techniques to analyse
large amounts of information to identify patterns, trends and other use-
ful information. TDM is used for training AI systems in order to perform
tasks in journalism, marketing, business analytics and by cultural institu-
tions. TDM can be understood as a research technique to collect infor-
mation from large amounts of digital data through automated software
tools. In the process of TDM, the analysed material is first being identi-
fied and data is individually collected or organised in pre-existing data-
bases. After that, substantial quantities of material are being copied.
These materials are first turned into a machine-readable format compat-
ible with a technology to be deployed for the TDM so that structured
data can be extracted and pre-processed material is possibly uploaded on
a platform, depending on a TDM technique to be deployed. Subsequently,
data is being extracted and recombined to identify patterns into the
final output.
98 L. Klobucnik
It is worth recalling that the UK was the first country in Europe to

have a TDM exception – limited text and data analysis exception was
introduced in S. 29A CDPA taking advantage of the possibilities under
Article 5(3)(a) of the InfoSoc Directive. The introduction of the TDM
exception was essential to unlock the potential of the British research and
unburden researchers from encumbrances and uncertainties. The current
exception covers only non-commercial purposes. TDM usually requires
copying of the material to be analysed. Therefore, in particular, the
infringement of reproduction as an act restricted by copyright pursuant
to the CDPA comes into question (section 17 CDPA). TDM techniques
open up tensions between rights and freedoms of text and data miners on
the one hand and exclusive rights of copyright holders on the other hand.
The (potentially) new regime of TDM exceptions should be mindful of
the balance of interests of these two groups. Reliance on the current
exceptions to the reproduction right do not provide a sufficient certainty
for TDM. Currently, exceptions to reproduction right based on S.29A
CDPA only cover temporary, transient and incidental reproductions with
no economic significance. Especially the last requirement is hard to sat-
isfy. Thus, the risk of infringing copyright when performing TDM activi-
ties is relatively high. The intention of the proposed exception is that
rights holders will no longer be able to charge for UK licences for TDM
and will not be able to contract or opt out of the exception. The new
provision may also affect those who have built partial business models
around data licensing. The main safeguard for rights holders will be the
requirement of lawful access. That means that rights holders could choose
the platform where they make their works available, including charging
for access via subscription or single charge. They will also be able to take
measures to ensure the integrity and security of their systems.
3.2.2 TDM Exception Not Allowing Rights Holders

to Opt Out
The UK Government intends that anyone with lawful access to material

protected by copyright should be able to carry out this analysis without
further permission from the copyright owner. Among other uses, data
mining can be used when training AI systems. For example, machine-

learning software which has been trained on large repositories of com-
puter code is able to intelligently suggest new code to programmers.
Although the government claims that the suggested TDM exception will
bring benefits to a wide range of stakeholders in the UK, including
researchers, AI developers, and cultural heritage institutions, the biggest
concern of copyright holders may be that they will not be able to control
who has access to their work. While the TDM exception may encourage
innovation in AI technology and promote its use for the public good, it
should also aim to preserve the central role of AI in promoting human
creativity. Under the proposed TDM exception, rights holders will not be
allowed to opt out of text and data mining AI systems. This proposal has
already attracted some criticism. During the stakeholder response phase
to the UK Government’s AI and IP Consultation, music rights holders
expressed their concerns that music would be reduced to mere data for
the purposes of mining. The proposal does not seem to take into account
that datasets are present in all types of works, not only in databases. It was
argued that the non-existence of the opt out regime essentially removes
any opportunity for creators to get compensation for their creativity and
talent, and they will also not be able to prevent the abuse of their work by
a computer. Some right holders stated their preference to remove their
works from datasets used to train AI systems.
3.2.3 Potential New TDM regime
The proposed TDM regime should promote a balance between all stake-
holders involved in the copyright value chain – rights holders, users, and
consumers. TDM is not only important for non-commercial research,
but also for other purposes, such as freedom of speech and journalism.
Therefore, allowing access of a large variety of entities to TDM is essen-
tial. Most importantly, TDM is a tool for AI-enabled innovation. Much
of the current and future development of AI depends on TDM. Therefore,
it is important to ensure that no entities are excluded from a possibility
to perform TDM activities. For instance, start-ups, which are not affili-
ated with any research institution and operate on a for-profit basis,
100 L. Klobucnik
cannot currently avail themselves of the TDM exception and can get
caught up in robust licensing mechanisms which can prevent start-ups
from emerging.
Researchers are faced with legal uncertainty as to whether and under
which conditions can they carry out TDM on content they have lawful
access to. TDM exceptions can and should go further in adapting to the
digital environment. It has to be noted that the TDM exception is also
dealt with in the EU Copyright in the Digital Single Market Directive
2019/ 790 (EU CDSM Directive). Unlike the current UK TDM excep-
tion, the TDM exception pursuant to the EU CDSM Directive applies
to both commercial and non-commercial mining purposes. The differ-
ence between the two being that if mining is performed for non-
commercial purposes (i. e. by research organisations or cultural
institutions), right holders do not have a right to opt out of the TDM
exception. If TDM is done for commercial purposes, right holders are
allowed to opt out, i. e. to contractually override the exception to the
reproduction right infringement. The currently applicable UK regime of
TDM exceptions does not distinguish between mining for commercial or
non-commercial purposes. The newly proposed TDM exception blurs
the distinction between TDM done for commercial or non-commercial
purposes and does not allow rights holders to opt out.
Copyright law has to respond to the realities of the twenty-first cen-
tury where the growing use of big data and AI tools in research and inno-
vation are necessary to achieve breakthroughs and result from collaboration
of different stakeholders – start-ups, SMEs, academia, research groups,
governments and business. It should be noted that the main beneficiary
of innovative projects is the public. In order for the public to benefit,
TDM exceptions should exist for all entities and for all purposes. If the
opt out regime is in place, rights holders have control over TDM – they
can decide who to grant a licence to or to entirely prohibit the use of their
content for TDM. For instance, publishers, who already offer paid-for
TDM as value-added services, will be reluctant to grant TDM licences to
third parties. Thus, commercial AI developers, journalists, research hubs,
and other innovative entities will be at a competitive disadvantage. The
opt out regime might cripple innovation for wide range of market play-
ers, from large companies to start-ups and individual researchers with a
particular emphasis on the game changing field of AI-based innovation.

Another downside of the opt out regime is that only a few research organ-
isations will be able to acquire licences for all copyright protected content
that is relevant for TDM research. Moreover, comprehensive TDM
research projects would be impossible to perform for the majority of
research organisations, especially if they have limited access to funding.
The opt out mechanism strongly limits the effectiveness of the TDM
exception and its ability to promote competitive advantage for compa-
nies engaged with AI. The opt out regime would not be able to prevent a
situation where dominant market players customarily override exceptions
by imposing both contractual and technological measures, thus depriving
users of the enjoyment of exceptions and lawful uses. An alternative to an
opt out regime would be a fair-remuneration mechanism, implemented
by way of an amendment to the CDPA and operated, for instance, by the
UK Intellectual Property Office.
4 DABUS Case – A Milestone in Patent Case

Law or an Unsuccessful AI Challenge
to Patent Law
4.1 AI and Patent Law
Copyright law is not the only area of IP that seems to be relevant with
respect of involvement of AI in the creative or inventive process. Patent
offices and courts in several countries, including the United Kingdom,
have been dealing with the question as to whether artificial intelligence
(and thus a non-human) can be an owner of a patent or considered an
inventor. Patent offices and courts in several countries have provided – at
least tentatively – answers to these questions.
4.2 United Kingdom
In 2018, a team of patent attorneys acting on behalf of an AI scientist,

Dr. Stephen Thaler, submitted two patent applications listing an AI
102 L. Klobucnik
system, DABUS, as an inventor. One of the patent applications was for a

type of a food container and one for a flashing light. However, Dr. Thaler
did not list himself as the inventor. Instead, he listed the DABUS AI
system, arguing that DABUS should be granted the patent ‘by ownership
of the creativity of the machine’ – while making clear that DABUS was
the inventor, not Mr. Thaler. The UK Intellectual Property Office
(UKIPO) told Mr. Thaler that he needed to list a real person as an inven-
tor. Since he did not do so, the UKIPO considered the application as
withdrawn and refrained from any further steps. Dr. Thaler claimed that
DABUS, which stands for ‘Device for Autonomous Bootstrapping of
Unified Sentience’, invented the flashing light and the food container
based on fractal geometry that were listed in the disputed patent applica-
tions. The UK was the first country where the patent office’s decision was
reviewed by a court. On 21 September 2021, Justice Marcus Smith in the
High Court of England and Wales dismissed Dr. Thaler’s appeal and
upheld the UKIPO decision, confirming that AI cannot be listed as an
inventor because it is not a person.
The core argument of the High Court was whether a law written for
human inventors can be applied to machines. The court ruled against Dr.
Thaler. Dr. Thaler even claimed that the artificial inventor DABUS could
appreciate its creations, is equipped with learning rules to bind memo-
ries, contained with a series of nets, together to produce not only com-
plex concepts, but also the consequences of said concepts, what
psychologists would call affective responses. In other words, feelings or
sentience was the result.
More recently, the England and Wales Court of Appeals (EWCA) in
the case Stephen Thaler and Comptroller General of Patents and
Trademarks and Designs from 21 September 2021 by majority (Arnold
LJ and Laing LJ; Birss LJ dissenting) upheld the judgment at first instance.
Dr. Thaler argued before the EWCA that the definition of the inventor as
the ‘actual deviser’ of an invention pointedly did not refer to the need for
the inventor to be a person. While the Patents Act 1977 was clearly writ-
ten on the assumption that inventors were persons, that was not, an
ought not to be, a requirement of the law. The Court of Appeal held that
an inventor must be a real human person under the UK law. Lady Justice
Elisabeth Laing of the Court of Appeals wrote in her judgment that ‘only
a person can have rights, a machine cannot’. Patent is a statutory right

and can only be granted to a person. Thus, this case was ultimately not
about whether an invention is patentable on its own, but rather about
whether a machine can have rights. According to the Court of Appeal, it
cannot have rights arising out of patent, because it cannot have any other
rights. Lord Justice Arnold agreed with the view that according to the
Patents Ac 1977, only a person can be an inventor. However, the third
Judge, Lord Justice Birss, took a different view in arguing that while
machines are not persons, the law did not demand a person to be named
as the inventor at all. In his opinion, the fact that in a patent application
there is a possibility for not including any inventor, simply means that
the Intellectual Property Office does not have to name anyone as an
inventor. In his opinion, the UKIPO is not obliged to name anyone or
anything as an inventor. While the three judges agreed that AI cannot be
listed as an inventor, their opinions differed on whether the UKIPO has
sufficient grounds to consider the application as withdrawn. Justice Arnold
pointed out that it is not for courts to weigh policy options and to debate
what the law ought to be. If Mr. Thaler listed himself as the inventor,
such an application would be successful. On the 20th of December 2023,
the UK Supreme Court confirmed the view of the Court of Appeals, not-
ing that according to the currenly applicable law, a patent requires disclo-
sure of an inventor, who must be a natural person. Thus, in case of an
AI-generated invention without an identifiable human inventor, the
invention is unprotectable. The UK Supreme Court noted that if patents
are to be granted in respect of inventions made by machines, the 1977
Patents Act will have to be amended.
By the operation of the Patent Cooperation Treaty (PCT), the applica-
tions filed in the UKIPO were extended to a number of countries, includ-
ing the US, Germany, Europe, Australia, South Korea, Japan, Israel,
Canada, New Zealand, Taiwan and others. Thus far, patent applications
have been rejected not only by the UKIPO, but also by the European
Patent Office (EPO), German Patent and Trademark Office, US Patent
and Trademark Office (USPTO), and Australian Intellectual Property
Office (AUIPO). Dr. Thaler decided to appeal patent registration rejec-
tions from the relevant patent offices. Let us have a closer look at applica-
tions in other countries.
104 L. Klobucnik
4.3 USA
In the United States, the DABUS application was had been initially
rejected by the United States Patent and Trademark Office (USPTO),
which took the view that the definition of the ‘inventor’ in the Patent Act
requires an inventor to be a natural person. Consequently, the USPTO
rejected the patent application because no natural person was listed as an
inventor. Dr. Thaler appealed the USPTO’s decision. On appeal, in
September 2021, Judge Leonie Brinkema in the United States District
Court for the Eastern District of Virginia upheld the decision of the
USPTO that an AI machine cannot be an inventor under the US patent
law. In reaching this decision, Brinkema J pointed out that the language
used in the Patent Act (35 U.S. Code) and the recent America Invents
Act 2011 refers to a human quality, especially in terms such as ‘individ-
ual’, ‘himself or herself ’. According to Brinkema J, this implies that an
inventor must be a human being. The decision was appealed by Dr.
Stephen Thaler, but the Supreme Court of the United States declined to
hear the appeal in April 2023.
4.4 Germany
In line with the reasoning adopted in the UK and the USA, in Germany,
the Examining Section of the German Patent and Trade Mark Office
(Deutsches Patent- und Markenamt, hereinafter DPMA), rejected the
patent application filed by Dr. Thaler on the grounds that according to
Section 6 and Section 37 of the Patent Act an inventor can only be a
natural person – a human being with legal capacity under Section 1 of
the Civil Code – which does not apply to AI. The crux of the legal prob-
lem was whether only a natural person can be named as in inventor
within the meaning of Section 37 of the Patent Act. The Federal Patent
Court noted in line with courts in other jurisdictions, that only a natural
person can be designated as an inventor in a patent application. This
conclusion was reached based on a doctrinal understanding of the con-
cept of the ‘inventor’ and a teleological interpretation of the respective
provisions under the German Patent Law. The German Federal Patent
Court viewed the right to be named as an inventor as a personality right,

akin to moral rights in copyright. It pointed out that the inventor’s right
to be named under Sections 37 (1) and 63 of the Patent Act is above all a
moral right, and such right can only be borne by humans. The Board
stated that the existing provisions reflect ‘the decision taken deliberately
by the legislature’, from which it deduced that ‘an AI can never be desig-
nated as an inventor or a co-inventor’.
The judgment relied on one of the underlying notions of patent law
that an invention is the result of human creativity or ingenuity. However,
this seems to contradict the role of Dr. Thaler in the inventive process.
Dr. Thaler claimed that DABUS devised the invention at issue com-
pletely autonomously and on its own, and that Dr. Thaler had no influ-
ence on the task and its solution, which led to the present invention. By
the same principle as DABUS was rejected as an inventor, patent offices
and courts should have prevented an outcome where Dr. Thaler was des-
ignated as the inventor and could be entitled to inventor’s rights. The
German Federal Patent Court held that the designation of inventorship
is an expression of the inventor’s personality rights. DABUS can be men-
tioned in the ‘inventor designation’, but only as long as Dr. Thaler’s name
is mentioned, too, and if DABUS has a subsidiary role. There is an irony
in the fact that Dr. Thaler became entitled to such rights against his own
testimony of not having created an invention himself and performing a
rather organisational activity.
The judgment of the Federal Patent Court on the one hand reaffirmed
the principle that human creativity underlines the inventor’s right to be
named and on the other hand, it allowed someone who, in his own words,
had no influence on the task and its solution to be named as an inventor,
which led to the present invention. If the principle of human creativity
was applied consistently and if Dr. Thaler’s claim regarding the genesis of
the invention had been treated as a statement of fact, rather than an inter-
pretation of the law, the case outcome would be that no one could be
designated as the inventor, and, hence, no one could potentially be enti-
tled to the patent for the food container.
106 L. Klobucnik
4.5 South Africa
The application to register DABUS as an inventor was also made in South

Africa. The South African Companies and Intellectual Property
Commission became the world’s first patent office to award a patent
which named AI as the inventor of a product. It should, however, be
noted that South Africa has a less rigid patent examination procedure – it
does not presently conduct substantive examination of patent applica-
tions in the way that the EU, the UK, Australia and the US do.
4.6 Australia
In the light of the aforementioned, the decision of the Federal Court of

Australia (FCA) in Thaler v Patent Commissioner seems to be an interna-
tional outlier. On 30 July 2021, Justice Beach overturned the primary
decision of IP Australia and concluded that AI could be listed as an inven-
tor under the Patent’s Act. It is the first court decision in the world sug-
gesting that AI can be an inventor under the current patent law. Beach J
emphasised that the patent application in question suggests AI as an
inventor, not as a patentee. Understandably, since AI does not have a
legal personality and cannot be an owner of any rights, it cannot be an
owner of a patent. He further emphasised that the understanding of a
patentee should clearly be distinguished from an inventor. While only a
human or other legal person can be an owner, controller or a patentee, an
inventor may be an artificial intelligence system. However, in these cir-
cumstances the inventor cannot be the owner, controller or patentee of
the patentable invention. An interesting observation was made by Beach
J with respect to AI being involved in other areas of intellectual prop-
erty law. In his opinion, patent law does not require a human ‘creator’ –
this is however, not true with respect to copyright law, because the
existence of moral rights in copyright law implies that only a human can
be considered an author for the purposes of copyright law. Unlike a pat-
ented invention, a copyright-protected work originates by the applica-
tion of human intellectual effort. Indeed, as discussed in this paper, the
concept of authorship would be not be fitting for creations by artificial
intelligence, but copyright law has other vehicles to accommodate such

creations.
However, this unique and rather controversial decision was reversed in
April 2022 by the Federal Court of Australia. The panel of judges pointed
out that courts should be cautious about interpreting statutes by refer-
ence to what might be considered desirable policy to avoid imputing that
policy into legislation and characterising that as the purpose of the legis-
lation. The Court respectfully noted that Dr. Thaler intended to provoke
a debate about the extent to which an AI may be an inventor. However,
it also suggested that this policy argument “clouded” the question before
Beach J.
5 Conclusion
The analysis above shows that despite AI being at the forefront of IP dis-
cussions in the past years, legislators, courts, and policy makers are not
quite ready yet to accept AI as creators or inventors. The understanding
that there is inevitably a ‘human behind AI’, without whom a subsequent
creation or invention by AI would not be possible, seems to prevail.
Granting IP rights to AI would raise a host of further legal questions
beyond mere IP rights, for instance questions of legal personhood. The
question of machines obtaining IP protection should ultimately be
decided by policy-makers. Some changes in law and policy can be antici-
pated, since it is becoming increasingly difficult for humans to discern
whether a song or a painting was created or a device was invented by a
machine. Ultimately, a policy decision will have to be made as to what
type of IP protection should be given to creations and inventions devised
by an algorithm with little or no human intervention. In the area of pat-
ent, developments may be slow, as it seems that statutory amendments
will be needed in order to grant patent rights to AI or recognise AI as an
inventor. The latter may be plausible. According to the AIPPI resolution,
AI inventions should not be excluded from patent protection per se. At
the same time, AI inventions should only be granted patent protection if
a natural person is listed as an inventor or co-inventor. However, the
input of a natural person can be quite minimal – even limited to the
108 L. Klobucnik
recognition that an output of an AI algorithm constitutes an invention.

The UK Government currently holds the position that no change to the
UK patent law is being considered because AI is not yet advanced enough
to invent without human intervention. However, at the same time, the
UK Government noted that the area of patent law will have to be kept
under review.
The area of copyright law seems to offer a few options to respond to
the challenge of artworks created by AI. This would be particularly by
granting AI rights akin to neighbouring (related) rights to copyright,
since these rights were created for entities not involved in the creative
process, e. g. rights of phonogram producers. Moreover, copyright law
has already shown a certain level of flexibility by granting protection to
works where there is no human author, i. e. computer-generated works.
It seems almost certain that legislators and policy makers will respond to
AI challenges to copyright. However, such response will most likely not
have a form of a major disruption of IP rights, but rather an incremental
change in the form of new rights tailored for AI.
References
AIPPI – Association for the Protection of Intellectual Property, ‘Copyright and
Artificially Generated Works, Resolution of the 2019 AIPPI World
Congress – London, 18 September 2019’ <http://114.247.84.87/AIPPI/
ztyj/jy/201911/P020191219538152201660.pdf>
——, ‘Inventorship of Inventions Made Using Artificial Intelligence, Resolution
of the 2020 AIPPI World Congress’ <https://aippi.nl/wp-content/
uploads/2020/10/Adopted-Resolution-Q 272-Patents-Inventorship-o f-
inventions-made-using-Artificial-Intelligence.pdf>
Aplin T and Davies J, Intellectual Property Law: Text, Cases, and Materials
(Fourth Edition, Oxford University Press)
Berne Convention for the Protection of Literary and Artistic Works 29
Čerka P, Grigienė J and Sirbikytė G, ‘Liability for Damages Caused by Artificial
Intelligence’ (2015) 31 Computer Law & Security Review 376
Drexl J and others, ‘Artificial Intelligence and Intellectual Property Law –
Position Statement of the Max Planck Institute for Innovation and
Competition of 9 April 2021 on the Current Debate’ (9 April 2021) <https://
papers.ssrn.com/abstract=3822924> accessed 15 December 2022
——, ‘Technical Aspects of Artificial Intelligence: An Understanding from an

Intellectual Property Law Perspective’ Max Planck Institute for Innovation
and Competition Research Paper Nr. 19 – 13 15
European Parliament, ‘Report with Recommendations to the Commission on
Civil Law Rules on Robotics’ <https://www.europarl.europa.eu/doceo/
document/A-8-2017-0005_EN.html> accessed 6 July 2020
‘Founder’ <https://imagination-engines.com/founder.html> accessed 2
December 2022
Fried I, ‘Exclusive: Adobe Will Sell AI-Made Stock Images’ (Axios, 5 December
2022) <https://www.axios.com/2022/12/05/adobe-ai-made-stock-images>
accessed 6 December 2022
Geiger C, Frosio G and Bulayenko O, ‘Text and Data Mining: Articles 3 and 4
of the Directive 2019/790/Eu’ [2019] SSRN Electronic Journal <https://
www.ssrn.com/abstract=3470653> accessed 28 December 2021
Guadamuz A, ‘Do Androids Dream of Electric Copyright? Comparative
Analysis of Originality in Artificial Intelligence Generated Works’ (2017)
2017 Intellectual Property Quarterly 18
Hartmann C and others, ‘Trends and Developments to Artificial Intelligence:
Challenges to the Intellectual Property Rights Framework’ <https://op.
europa.eu/en/publication-d etail/-/ publication/394345a1-2 ecf-1 1eb-
b27b-01aa75ed71a1/language-en>
Harwell D, ‘He Used AI to Win a Fine-Arts Competition. Was It Cheating?’
(Washington Post, 2 September 2022) <https://www.washingtonpost.com/
technology/2022/09/02/midjourney-a rtificial-i ntelligence-s tate-f air-
colorado/> accessed 12 December 2022
Hilty RM, Hoffmann J and Scheuerer S, ‘Intellectual Property Justification for
Artificial Intelligence’ 29
Hugenholtz PB, ‘The New Copyright Directive: Text and Data Mining (Articles
3 and 4)’ (Kluwer Copyright Blog, 24 July 2019) <http://copyrightblog.klu-
weriplaw.com/2019/07/24/the-n ew-c opyright-d irective-t ext-a nd-d ata-
mining-articles-3-and-4/> accessed 27 December 2021
Kelly C, ‘Australian Artists Accuse Popular AI Imaging App of Stealing Content,
Call for Stricter Copyright Laws’ The Guardian (11 December 2022)
<https://www.theguardian.com/australia-news/2022/dec/12/australian-
artists-accuse-popular-ai-imaging-app-of-stealing-content-call-for-stricter-
copyright-laws> accessed 12 December 2022
Kim D, ‘The Paradox of the DABUS Judgment of the German Federal Patent
Court’ [2022] GRUR International ikac125
110 L. Klobucnik
Koempel F, ‘UK Music’s Legal Expert Outlines Issues With Proposed AI Data
Mining’ (UK Music, 29 November 2022) <https://www.ukmusic.org/news/
uk-m usics-l egal-e xpert-o utlines-i ssues-w ith-p roposed-a i-d ata-m ining/>
Matulionyte R, ‘AI as an Inventor: Has the Federal Court of Australia Erred in
DABUS?’ (2022) 13 JIPITEC <http://www.jipitec.eu/issues/jipitec-13-
2-2022/5538>
McCutcheon C, ‘Curing the Authorless Void: Protecting Computer-Generated
Works Following ICE TV and Phone Directories’ (2013) 37 Melbourne
University Law Review
McDermott E, ‘DABUS Scores Again with Win on AI Inventorship Question
in Australia Court’ (IPWatchdog.com | Patents & Intellectual Property Law,
2 August 2021) <https://ipwatchdog.com/2021/08/02/dabus-scores-win-ai-
inventorship-question-australia-court/id=136304/> accessed 25 November
2022
Metz R, ‘These Artists Found out Their Work Was Used to Train AI. Now
They’re Furious | CNN Business’ (CNN, 21 October 2022) <https://www.
cnn.com/2022/10/21/tech/artists-ai-images/index.html> accessed 8
December 2022
‘Patents and Applications – The Artificial Inventor Project’ <https://artificialin-
ventor.com/patent-applications/> accessed 15 December 2022
Ramalho A, ‘Will Robots Rule the (Artistic) World? A Proposed Model for the
Legal Status of Creations by Artificial Intelligence Systems’ (2017) 21 The
Journal of Intenet Law 12
———, Intellectual Property Protection for AI-Generated Creations: Europe,
United States, Australia, and Japan (Routledge 2021)
Ricketson S, ‘The 1992 Horace S. Manges Lecture – People or Machines: The
Berne Convention and the Changing Concept of Authorship’ 16 Colum-
VLA JL & Arts
Senftleben M and Buijtelaar L, ‘Robot Creativity: An Incentive-Based
Neighboring Rights Approach’ [2020] SSRN Electronic Journal <https://
www.ssrn.com/abstract=3707741> accessed 17 November 2022
Stassen M, ‘Over 1,000 Songs with Human-Mimicking AI Vocals Have Been
Released by Tencent Music in China. One of Them Has 100m Streams.’
(Music Business Worldwide, 15 November 2022) <https://www.musicbusi-
nessworldwide.com/over-1000-songs-human-mimicking-ai-vocals-have-
been-released-by-t encent-m usic-i n-c hina-o ne-o f-t hem-h as-over-1 00m-
streams/> accessed 12 December 2022
The Government of the United Kingdom, ‘Artificial Intelligence and Intellectual

Property: Copyright and Patents: Government Response to Consultation’
(GOV.UK) <https://www.gov.uk/government/consultations/artificial-
intelligence-and-ip-copyright-and-patents/outcome/artificial-intelligence-
and-intellectual-property-copyright-and-patents-government-response-to-
consultation> accessed 8 December 2022
Trapova and Gervassis Ni, ‘UKIPO’s Public Consultation on AI and IP –
Computer-Generated Works (Part 1)’ (Kluwer Copyright Blog, 14 March
2022) <http://copyrightblog.kluweriplaw.com/2022/03/14/ukipos-public-
consultation-on-ai-and-ip-computer-generated-works-part-1/> accessed 15
December 2022
UK Intellectual Property Office, ‘Artificial Intelligence and Intellectual Property:
Copyright and Patents: Government Response to Consultation’ (GOV.UK)
<https://www.gov.uk/government/consultations/artificial-intelligence-and-
ip-copyright-and-patents/outcome/artificial-intelligence-and-intellectual-
property-copyright-and-patents-government-response-to-consultation>
———, ‘Government Response to Call for Views on Artificial Intelligence and
Intellectual Property’ (GOV.UK) <https://www.gov.uk/government/consul-
tations/artificial-i ntelligence-a nd-i ntellectual-p roperty-c all-f or-v iews/
government-r esponse-t o-c all-f or-v iews-o n-a rtificial-i ntelligence-a nd-
intellectual-property> accessed 27 June 2021
van Gompel S, ‘Creativity, Autonomy and Personal Touch: A Critical Appraisal
of the CJEU’s Originality Test for Copyright’ in Mireille Eechoud (ed), The
Work of Authorship (Amsterdam University Press 2014)
Vanpoucke W, ‘Copyright Challenged by Art Created by Artificial Intelligence’
(2021) 43 European Intellectual Property Review 12
Vincent J, ‘Getty Images Bans AI-Generated Content over Fears of Legal
Challenges’ (The Verge, 21 September 2022) <https://www.theverge.
com/2022/9/21/23364696/getty-i mages-a i-b an-g enerated-a rtwork-
illustration-copyright> accessed 2 December 2022
Wiggers K, ‘GitHub Launches Copilot for Business Plan as Legal Questions
Remain Unresolved’ (TechCrunch, 8 December 2022) <https://techcrunch.
com/2022/12/08/github-l aunches-c opilot-f or-b usiness-p lan-a s-l egal-
questions-remain-unresolved/> accessed 12 December 2022
World Intellectual Property Organisation (WIPO), ‘The WIPO Conversation
on Intellectual Property and Artificial Intelligence’ <https://www.wipo.int/
112 L. Klobucnik
about-ip/en/artificial_intelligence/conversation.html> accessed 11
December 2022
Commissioner of Patents v Thaler (Australia) [2022] Federal Court of
Australia FCAFC 62
Decision on patent application GB18169094 and GB18181610, BL O/741/19
[2019] Intellectual Property Office of the United Kingdom BL O/741/19
Dr Stephen L Thaler v Presidentin des Deutschen Patent- und Markenamts,
1 W (pat) 5/21 (Bundespatentgericht (Germany))
Eva Maria Painer v Standard Verlags GmbH [2011] Court of Justice of the
European Union ECLI:EU:C:2011:798
Infopaq International A/S v Danske Dagbladens Forening, Case C-5/08,
ECLI:EU:C:2009:465 [2009] Court of Justice of the European Union
ECLI:EU:C:2009:465
Stephen L Thaler v The Comptroller-General of Patents, Designs and Trade
Marks [2020] EWHC 2412 (Pat) [2020]
Thaler v Commissioner of Patents [2021] Federal Court of Australia FCA 879
Thaler v Comptroller General of Patents Trade Marks And Designs [2021]
EWCA Civ 1374 (EWCA (Civ))
Thaler v Iancu, et al, No 1:2020cv00903 – Document 33 (ED Va 2021) [2021]
United States District Court for the Eastern District of Virginia
1:2020cv00903
Copyright, Designs and Patents Act 1988 (United Kingdom) 1988
Directive 2001/29/EC of the European Parliament and of the Council of 22
May 2001 on the harmonisation of certain aspects of copyright and related
rights in the information society, OJ L 167
Directive 2019/790/EU of the European Parliament and of the Council of 17
April 2019 on copyright and related rights in the Digital Single Market and
amending Directives 96/9/EC and 2001/29/EC, OJ L 130/92 2019
H.R.1249 – Leahy-Smith America Invents Act
Patent Act 1990 (Act No. 83) (Australia)
Patentgesetz (Germany) – Patent Act as published on 16 December 1980
(Federal Law Gazette 1981 I p. 1), as last amended by Article 4 of the Act of
8 October 2017 (Federal Law Gazette I p. 3546)
Patents Act 1977 (United Kingdom) 1977
Judgment of the Supreme Court, Thaler (Appellant) v Comptroller-General of
Patents, Designs and Trade Marks (Respondent) [2023] UKSC 49
An Artificial Intelligence Invention
Protection Model
Budi Agus Riswandi
1 Introduction
The industrial revolution period reflects the advancement of technology
in human civilization. All the industrial revolutions that have occurred
have had an impact on economic growth, increased productivity, and
high-quality goods and services (Rabeh Morrar, 2017). The industrial
revolution is currently in the 4.0 phase which is marked by the prolifera-
tion of technology based on computer programs that perform their func-
tions more like a human’s ability to think and act, resulting in the
technology known as artificial intelligence technology.
The terminology of AI was formally coined by John McCarthy, a com-
puter scientist at a conference in 1956, according to him, it was the
notion of a program, processing and acting on information, such that the
result is parallel to how an intelligent person would respond in response
to similar input (Tripathi & Ghatak, 2018). In short, the term ‘Artificial
B. A. Riswandi (*)
Faculty of Law, Universitas Islam Indonesia, Yogyakarta, Indonesia
e-mail: budiagusr@uii.ac.id
https://doi.org/10.1007/978-3-031-42576-9_5
114 B. A. Riswandi
Intelligence’ (hereinafter referred to as AI) refers to those computer sys-

tems capable of performing tasks that would normally require some intel-
ligence if done by humans (Intellectual Property Office, 2019).
Along with the industrial revolution era, there was a society era that
reflected human social behavior in the use of technological innovations.
Society is currently in its 5.0 chapter, which is marked by science and
technology by combining virtual and real space in the form of big data,
internet of things, and AI in assisting with various human activities.
According to the Japanese prime minister, Shinzo Abe said that the con-
cepts of industrial revolution 4.0 and society 5.0 do not have much dif-
ference. namely, industrial revolution 4.0 uses AI while society 5.0 focuses
on the human component.
AI is a general-purpose technology increasingly present in all aspects of
our lives worldwide and across all industry segments. Thus, it cannot be
denied that AI is a transformative technology, which has revolutionized
many areas of human life. While the technology goes back to the 1950s,
the recent exponential increase in data and computing power and
improved mathematical models are largely responsible for the current
widespread implementation of AI. AI technologies are starting to perme-
ate creative and innovative activities, which until recently were perceived
to be solely the remit of humans. The current intellectual property (IP)
system was designed to incentivize that human creation and innovation.
As AI is operating more autonomously, this raises fundamental questions
for the IP system across all IP rights (World Intellectual Property
Organization, 2021).
The development of technology in general and AI has led to funda-
mental changes in the invention process (Simon, 2013). Various exam-
ples of currently available and easily recognizable technologies AI
technology include Search Engines for websites or e-commerce platforms,
Chat Bots, Voice assistants such as Siri for iOS, and Google Assistant,
Virtual Travel Assistants, Image Recognition, Humanoid/Robotics such
as Tesla Bot, iRobot, and Self-Driving Car. The example above of AI
technology can help solve human problems in cyberspace and real space
by enabling the combination of software and hardware. Moreover, AI is
also being developed to help solve some the world’s biggest challenge:
starting from the treatment of chronic diseases or reducing the death rate
An Artificial Intelligence Invention Protection Model 115
in traffic accidents, to mitigating climate change or tackling security

threats in cyberspace.
However, behind its function in displaying intelligent behavior, AI
also has the potential to trigger certain legal problems in certain situa-
tions. One of them is related to intellectual property law. Few people pay
attention to the challenges that AI presents to intellectual property legal
frameworks. Eventhough, the growing importance of AI technology and
the gaps in both the copyright and patent systems identified by experts
clearly point to the need for further investigation (Burt, 2021).
AI would be defined as the simulation of human intelligence on a
machine in order to make the machine efficient to identifying and using
the right piece of ‘knowledge’ at a given step of solving a problem. AI has
the ability to think and act like humans, and its complex understanding
is a computer program that receives various data as input and then pro-
cesses it using an algorithm as a method of solving problems using special
techniques in the form of minimal machine learning or deep learning to
process input data and output results in the form of software that can be
operated with an operating system or with hardware whose actions
resemble human capabilities.
AI refers to computer programs that can solve problems that previ-
ously required human-level programming and coding expertise. AI dif-
fers from traditional rule-based programs in that it does not always
require a programmer to code all of the rules, instead, it can figure out
which rules to use to produce the desired result. This process delegated
the heavy lifting to the algorithm, which saved time while also allowing
these programs to solve complex problems that would be impractical or
impossible to code using the traditional rule-based system (Reinaldo
Franqui Machin, 2021). AI can be defined simply as computer program
technology. in other words, AI is a form of the recent advancement in
computer program technology that exhibits human-like abilities in think-
ing and acting.
A computer program’s main elements are Algorithms, Programming
Languages, and Programs, all of which are also owned by AI technology
too. The input data is accompanied by special effects or techniques in
order for the algorithm to have capabilities such as the way the human
brain thinks, and the final output is in the form of software produced by
116 B. A. Riswandi
the algorithm, which can provide an action similar to humans in cyber-

space and/or realspace with specific/compatible hardware.
In Indonesia, Computer programs are protected by two types of intel-
lectual property laws: namely copyrights and patents. In copyright law, a
computer program is referred to as a work provided that the computer
program is a set of instructions expressed in the form of language, code,
schematics, or in any form intended for the computer to work and can
only perform certain functions without special effects or techniques and
have no connection with the industrial world. Whereas in patent law, a
computer program can be referred to as an invention if the computer
program has special effects and techniques in carrying out its function in
solving a problem and has a relationship with the industrial world.
Trade-Related Aspects of Intellectual Property Rights (TRIPs) in
Article 27 paragraph 1 states that objects that can be patented are any
inventions in the form of products or processes in all fields of technology,
as long as the invention or technology is still new, contains inventive
steps and can be applied in industry.
A patent is an exclusive right granted by the state to an inventor for his
invention in the field of technology for a certain period to carry out the
invention himself or to give approval to another party to implement it. A
patent is the legal right of an inventor to exclude others from making or
using a particular invention. This right is sometimes termed an “intel-
lectual property right” and is viewed as an incentive for innova-
tion (Bronwyn H Hall, 2007). An invention is an inventor’s idea that is
funneled into a specific problem-solving activity in the field of technol-
ogy, such as the creation of a product or process innovation, or the
improvement and development of a product or process. Patent protec-
tion is limited the given jurisdiction and is territorial, which means that
the patent only protects what is claimed and in the area where the patent
is granted which means that patents only provides jurisdictional and
regional protection for the desired subject matter in the country or region
where the patent is issued (DJKI, 2019).
Patents are state-granted exclusive rights granted to inventors for their
technological innovations for a set period of time, after which the creator
may use the creation themselves or allow third parties to use it.
Furthermore, an innovation is the result of an inventor’s idea being
applied to a specific technological problem-solving activity, such as the

development and enhancement of a product or method.
When AI technology evolves into a sophisticated computer program
with a function in solving a problem and a connection to the industrial
world, patent law is the appropriate intellectual property law to protect
AI inventions. AI inventions are expected to be able to understand
unstructured data, reason, learn automatically, and be used to automate
the invention process (Blok, 2017).
Aristotelis Tsirigos as WIPO expert in the fields of artificial intelli-
gence, data, intellectual property rights, policy and innovation and
Director of the Applied Bioinformatics Laboratories at NYU School of
Medicine, in which Aristotelis explained that Artificial Intelligence is a
collective computer algorithm that slowly acquires basic human-like abil-
ities in the form of sight, speech, and navigation (World Intellectual
Property Organization, 2019).
WIPO gives the characteristics of AI technology, namely having mini-
mal technical effects embedded in the algorithm, the effects of these tech-
niques include; (1) fuzzy logic; (2) probabilistic reasoning; (3) machine
learning (classification and regression, support vector machine, neural
networks, deep learning, logical and relational learning, probabilistic
graphical models, rule learning, instance-based learning, lanent represen-
tation, bio-inspired approaches, supervises learning, unsupervised learn-
ing, reinforced learning, multi-task learning); (4) ontology engineering;
and (5) logic programming (description logistics & expert system).
Based on the above description, AI is classified as an object that can be
protected by patent law. However, how does the Indonesian Patent Law
System regulate AI as an invention that can be legally protected? Therefore,
it would be investigated the protection of AI Inventions in Indonesian
Patent Law and compared the protection of AI inventions between
Indonesian, United States, and Japanese Patent Laws, as well as identified
an Artificial Intelligence Invention Protection Model.
To resolve these issues, normative legal research methods based on the
statute approach and comparative approach methods are used. It is hoped
that by using normative research methods with a variety of approaches,
they will be able to study and comprehend artificial intelligence inven-
tions based on Indonesian patent law and compare them to patent laws
in the United States and Japan.
118 B. A. Riswandi
2 Discussion
2.1 Legal Protection of AI Invention Based
on Indonesia Patent Law
Essentially, AI is the ability of computer programs to initiate to resemble

humans. Basically, and the AI is an innovation in the field of computer
programs. Under Law Number 13 of 2016 about Patent (hereafter
referred to as the Indonesia Patent Law), There is no specific mention or
clearly stated about AI technology or AI inventions as patent objects base
on Indonesia Patent law.
In general, the patent object is an invention in the field of technology
or what is known as an invention, according to Article 1 Paragraph 1 of
Indonesian Patent Law the scope of the type of invention is a product,
process, and refinement and development of that product or process.
Article 27 Paragraph 1 TRIPS also states that an object that can be pat-
ented is any invention, whether it takes the form of products or processes
in all fields of technology. and in Article 27 paragraphs 2 and 3 it does
not expressly provide exceptions to computer programs or business meth-
ods in general from patents and there is no further definition of ‘Invention’
and ‘technology’
In this case, AI is defined as computer programs, and according to the
Article 4 paragraph (d) of Indonesia Patent law, if a computer program
lacks character, technical effects, and problem solving, it is not an inven-
tion and cannot become a patent object. But, if the computer program
has a special character, technical effects, and problem solving, it becomes
an invention and can become a patent object that is protected by law.
AI is the most recent technology in the field of computer program
technology in this era. The distinguishes between ordinary computer pro-
grams with AI is that the instruction code used by AI is more complex in
writing the code, using algorithms written in a specific programming
language, so that the resulting program can solve problems like a human
mostly human thinking and acting are predictedable and unpredictable.
But an Ordinary computer programs only use codes that have or do not
have characters, technical effects, and problem solving and problem solv-
ing that are predictable.
AI as an invention of technology can be classified as patented products

because AI is a physical entity/object in the form of software that can be
used in a variety of existing operating systems or specific operating sys-
tems and hardware. AI was invented as part of a computer program
invention. A computer program that can be protected by a patent must
include programs with characters (an instructions written in a program-
ming language), technical effects, and functions to solve problems, both
tangible and intangible, which are inventions that can be patented base
on paragraph (d) of Indonesia Patent law. According to IP Module for
Patents Issued by Directorat General Intellectual Property of Indonesia in
2019, emphasized that computer program inventions are an example of
patent product.
AI as an invention in the field of technology can be classified as patent
process for algorithms own by the AI. the process of creating AI through
the use of algorithms written in a programming language. The Algorithms
written in a programming language have instructions, special technical
effects which become source code (input), and then problem solving
(output) in the form of human-like thinking and acting abilities. The
designed algorithm creates a computer program with human-like abili-
ties, enabling the algorithm be patented as a patent process. Because an
algorithm is a process/method for evolving a computer program’s
capability.
The AI which is a computer program has a problem-solving method in
the form of an algorithm owned by a computer program that manages
the input data given to produce a certain output and has a technical effect
that is embedded into the AI algorithm, for an example that AI technol-
ogy can solve financial & e-commerce problems. The business methods
owned by AI inventions can be divided into offline and online which
have an impact on financial business activities (Miyamoto, 2003).
According to Article 4 paragraph (c) number 3 and paragraph (d) of
Indonesia Patent Law, AI is an invention that can be granted a patent as
an patent product for the software or as the patent process for the algo-
rithm. AI is a computer program that has codes, characters, technical
effects, and problem-solving that uses complex/complicated algorithms
and is written in a specific programming language. Meanwhile, AI can be
patented, but it must meet the substantive requirements known as
120 B. A. Riswandi
patentability requirements of patents as well as trace prior art in the form

of novelty, inventive step, and industrial applicability.
If AI invention fulfills the patentability requirements for granting pat-
ents, which include new inventions, inventive steps, and the ability to be
applied in the industry, it is classified as the scope of patent protection
with a term protection time of 20 years. Further, if an AI invention is
only in the form of development, it is classified as a simple patent protec-
tion scope with a 10-year protection period where a simple patent is
granted based on an application and each application can only be filed for
one invention or one independent claim and cannot be a divisional/frac-
tional application is made, but a simple patent can be converted into a
patent if it has several inventions or several independent claims.
Thus, Indonesia’s Patent law currently does not clearly state that AI is
not a barrier for an inventor to apply for a patent for an invention.
Because patent examiners continue to examine using objective and sub-
jective requirements where the requirements are a basis that has been
regulated in patent law and make these requirements in examining pat-
ents against a cutting-edge invention as is the case with AI inventions to
be granted patents or otherwise.
2.2 Comparative Studies on Protection of Artificial

Intelligences Invention Between Indonesia,
United States, and Japan Patent Law
A comparative study of AI inventions as patent objects in the United

States ‘Patent Act (35 U.S. Code)’ and Japan ‘(特許法 Tokkyohō)’. The
two countries’ patent laws explicitly do not mention AI technology, then
searches for patent provisions using the term computer program, as pre-
viously explained, because AI technology is part of a computer program.
The United States Patent Law only lists computer programs on the
part of the patent fee (35 U.S.C.41 and AIA 14) without any explanation
regarding computer program inventions and does not mention AI tech-
nology as an object of the invention protected under United States patent
law. however, in Court Decision No.16/524,350 on April 27, 2020, in
the decision of the United States Patent and Trademark Office (USPTO)
it has been confirmed that AI technology cannot be an inventor in the

United States Patent Law.
In the Japanese Patent Law itself, the Patent Law places provisions for
inventions, especially for computer programs, in Art 2 Act No.109 of
2006, the provisions consist of 1 Article with 4 verses.
The United States Patent Law only mentions computer programs, the
Japanese Patent Law includes provisions for computer programs and pro-
vides an explanation in the contents of the article, and Indonesian patent
law also includes provisions for computer programs, although in the
Elucidation of Indonesia patent Law. The American Patent Examiner,
namely the U.S. The Patent and Trademark Office (USPTO) has issued
reference guidelines for examining patents for AI inventions. some defi-
nitions of AI in the USPTO guidelines are defined by the U.S. National
Institute of Standards and Technology (USPTO, 2020).
The Japanese Patent Examiner, namely the Japan Patent Office, has issued
a reference guide for patent examination of AI inventions in the form of a
report issued by the Patent Examination Department (Japan Patent Office,
2020). The Indonesian Patent Examiner, namely the Directorate General of
Intellectual Property, does not yet have a reference guide for patent examina-
tion of AI inventions and is still using the patent examiner module issued in
2019, which in the module does not yet discuss AI inventions.
Although the three countries have not explicitly listed AI inventions in
their patent laws, they already have AI inventions that have been pro-
tected by patents in their countries, particularly the United States and
Japan, which were the first to have protection for AI patent inventions,
and Indonesia followed suit in 2019 in terms of granting patents for AI
inventions, namely inventions with patent numbers IDS000002530 and
IDS000002530.
2.3 An Artificial Intelligence Invention

Protection Model
Indeed, AI can help society become modern and even bring about very
basic changes, however, its impact on the human condition is clearly
uncertain. Therefore, efforts are needed to respond to the various
122 B. A. Riswandi
challenges that arise because of the impact of AI intelligently and

humanely and take advantage of its positive functions and roles to create
a more just and equitable society. Therefore, not a few legal scholars then
try to identify a model of AI protection in inventions.
A study conducted by Petr M. Morhat in 2018, claims that AI does
not need to be protected through copyright and patent regimes (Rassolov
& Chubukova, 2022). To find out the status of legal relationships involv-
ing AI, there are at least several types of legal structures and models. First
is the machine-centric concept. based on this concept, AI is seen as the
legitimate author of the work created. The second is the anthropocentric
concept. According to this concept, AI is considered as a tool that humans
use in the process of creating intellectual property products. Third is the
work-for-hire concept. This concept reveals that AI acts as a “hired
worker” creating intellectual property products. Fourth is the hybrid
authorship concept, where AI based on this concept is considered as a
co-author of human-created intellectual property products. Fifth is the
concept of “disappearing” or zero authorship. This concept deals with
very difficult situations when the concepts being investigated seem to
overlap in different variations.
The Pathetic Dot theory proposed by Lawrence Lessig can be used in
providing a form of protection model for AI. Previously, Lawrence Lessig
used the Pathetic Dot theory in cyber law regarding cyberspace.
Pathetic Dot theory or other names New Chicago School theory is a
theory of socio-economic regulation, this theory discusses how individual
lives are governed by four forces: law, norms, market, and technical infra-
structure/design (architect). Moreover, in order to form a model of legal
protection for AI inventions, it is necessary to consider the legal elements
of a multi-level mechanism of responsibility and control over the devel-
opment of AI itself, where it is considered as a system that is intercon-
nected between technical control, technological control, standardization
(sanitary, information, technical, industry standards, etc.), creation of
harmonization or unification of technical and technological regulations,
as well as organizing a comprehensive “law platform compliance” AI in
the form of multilevel “responsibility and control regulators”.
Lawrence Lessig identifies four forces that limit an individual’s actions,
where the four forces represent the sum of what governs individual
actions, both directly and indirectly (ex-post & ex-ante). Law, which
serves as a sanction and/or provides protection, societal norms, the mar-
ket, which determines activity based on supply and demand, and archi-
tecture, which is a configuration of human behavior/action. Despite their
different functions and effects, the four forces work together to regulate
or limit the space for individual action/behavior. Norms are constrained
by the stigma imposed by the community, markets are constrained by the
prices they set, architecture is constrained by the physical burden it
imposes, and laws are constrained by the threats posed by the rules.
The four forces of the Pathetic Dot theory can be used to protect an
invention, especially the invention of AI as a protection model. The four
forces are as follows:
2.3.1 Patent Law as ‘Law’
As a form of positive law provided by a country to inventors and their

inventions. In this case, for AI protection, patent law is one of the laws
that govern technology, particularly inventions. In general, patent law
regulates and protects the inventor’s exclusive rights to a technological
invention. The legislature is working on the patent law provisions. Patent
law has guiding principles or principles that are followed and underpin
patent law arrangements, particularly in Indonesia with Patent Law
Number 13 of 2016.
2.3.2 Social Norm as ‘Norms’
Social norms are general habits or rules that guide behavior in a commu-
nity group and have regional boundaries. The limit of social norms is
behavior that is considered appropriate for a community group, which is
also known as social rules or social regulations. The values of norms that
society owns and believes in can provide an assessment of the usefulness
of a technological invention to society.
Technological advancements, particularly inventions, can also influ-
ence how society perceives a technological advancement by leveraging
124 B. A. Riswandi
social norms that apply in that environment, such as ways, habits, rules
of conduct, customs, laws, or fashion. Furthermore, the development of
AI with engineering characteristics similar to humans may lead to com-
munity disagreements about these inventions.
Thus, social norms control the regulation of AI inventions because the
effects and risks of technological developments, particularly AI inven-
tions, can affect social interaction in society because people will use,
enjoy, and rely on technology in the future.
2.3.3 Stakeholder as ‘Market’
Stakeholders in the form of investors, inventors, distributors, and consum-

ers of a technological invention have an indirect impact on the existence of
technology in society, particularly AI. In this case, stakeholders or markets
refer to how stakeholders in the business world submit requests (demand)
and offers (supply) for a product and/or service, which affect the price to be
determined and the form of a product and/or service offered customized
on request. Every trade transaction must have demand and supply.
Even though AI inventions are still in the development stage, and the arti-
ficial intelligence invention products offered are not widely enjoyed or used
by the general public, but rather by people with sufficient financial means, AI
are currently in higher supply than market demand. Because the demand is
dominated by investors, conglomerates, and governments interested in pro-
viding these AI technologies. Market forces, on the other hand, influence
inventors’ behavior in order for them to present up-to-date inventions.
2.3.4 Source Code of Computer Program as ‘Architecture’
A Programmers as individual or group which establish a program/soft-

ware, the programmers in good faith create and develop technological
inventions for human welfare by following a code of ethics for program-
mers. The architectural aspect referred to in AI inventions is the program
code also known as the source code. What is meant by code is a set of
computer instructions in the form of an algorithm written in a program-
ming language that functions to give work orders to a computer or a
device to perform certain tasks. As previously stated, artificial intelligence

inventions are a subset of computer program technology, which also
includes algorithms, programming languages, and programs.
Essentially, an AI is a computer program with unique technical com-
ponents, such that the role of the programmer who provides input
instructions (source code) to determine the output of AI becomes the
most fundamental force that determines an AI tech’s behavior / action.
A programmer has a code of ethics in developing computer technol-
ogy, one of the codes of ethics issued by the Association for Computing
Machinery (ACM), where the code of ethics has been
These four forces together regulate the space for movement-
development as well as become a protection for AI inventions. these four
powers are applied to be a protection against AI inventions and can also
be used to be aware of and avoid the negative effects of AI inventions that
can harm humans, especially harm society.
Every year, AI technology is improved so that it has capabilities that are
increasingly similar to the ideal abilities of a human (Code Program as
architecture). Inventors of AI are granted legal protection for their cre-
ations to be morally and economically exploited (Patent Law). The role
of the community in assessing the fairness of a technological invention
that is constantly evolving each year (Social Norms). Technology inves-
tors who collaborate with inventors to develop the most recent techno-
logical inventions (Market). Through this AI protection model, adapted
from Lawrence Lessig’s known as Pathetic Dot theory, it can establish a
harmonious relationship between law and technology, especially cutting-
edge technology, so that people can enjoy using it in a positive way while
the inventor of the technology receives protection for their rights.
3 Conclusion & Suggestion

3.1 Conclusion
Article 27 paragraph 1 of TRIPs states that the object of a patent is an

invention in all fields of technology so that cutting-edge inventions such
as artificial intelligence inventions are regulated and protected in the
126 B. A. Riswandi
national patent law whose provisions are guided by TRIPs. The AI inven-
tions are explicitly not included in the patent laws of Indonesia, the
United States, and Japan, but the three patent laws of these countries
include provisions regarding computer programs. The patent laws of the
three countries in practice have protected AI inventions under the patent
laws of their respective countries, although only the United States and
Japan have issued patent guidelines for AI inventions.
Lawrence Lessig’s Pathetic Dot/New Chicago School theory can be
used in creating a protection model for AI invention. The protection
model uses 4 powers as protectors, namely the Patent Law “Law” as the
protection provided by the State; Social Norms “Norm” as a protection
from the cultural aspects of the surrounding community; Market
“Market” protection from stakeholders in the business world, and Code
Program / Source Code “Architecture” as protection from programmers
who develop a computer program technology. these four powers are
applied to be a protection against AI and can also be used to be aware of
and avoid the negative effects of AI tech
3.2 Suggestion
(a) The renewal of the Indonesian National Patent Law in dealing with
the development of AI technology does not need to be carried out in
approximately 10 years. Because AI inventions themselves can grant
patents based on Law Number 13 of 2016 concerning Patents and
based on data on the PDKI-Indonesia website there are 2 inventions
based on AI that have been granted patents in 2019 and 2020.
(b) Directorate General Intellectual Property of Indonesia should imme-
diately issue a manual regarding AI artificial intelligence inventions
to provide education to patent examiners, academics, and the general
public about AI artificial intelligence inventions from the point of
view of Indonesian patent law.
References
Blok P. The Inventor’s New Tool: Artificial Intelligence-How Does it Fit the
European Patent System, European Intellectual Property Review, (2017).
Bronwyn H Hall. Patent and Patent Policy, Oxford Review of Economic
Policy, (2007).
DJKI. Modul Kekayaan Intelektual Bidang Paten, Direktorat Jenderal Kekayaan
Intelektual Kementerian Hukum & HAM R.I, 2019.
Ilya M. Rassolov and Svetlana G. Chubukova, “Artificial Intelligence and
Effective Governance: Legal Framework”, Kutafin Law Review Vol. 9 Issues.
2, (2022), p. 314.
Intellectual Property Office, Artificial Intelligence: a worldwide overview of AI
Patents and Patenting by the UK AI sector, A Report by the Economics,
Research and Evidence team at the Intellectual Property Office United
Kingdom, June (2019).
Japan Patent Office. Recent Trends in AI-Related Inventions. Report: Patent
Examination Departement, (July 2020).
Jeffrey A. Burt, “The Revolutionary Impact of Artificial Intelligence on the
Future of the Legal Profession”, Kutafin Law Review, Vol. 8 Issues. 3,
(2021), p. 400.
Rabeh Morrar. The Fourth Industrial Revolution (Industry 4.0): A Social
Innovation Perspective. Technology Innovation Management Review Vol 7
Issue 11, (November, 2017).
Reinaldo Franqui Machin, “It’s Time for the AI Patent: The Case for an Artificial
Intelligence Patent Category”, Research Gate, October 2021.
Simon B.M. Implications of Technological Advancement for Obviousness,
Michigan Telecommunications, and Technology Law Review, (2013).
Swapnil Tripathi & Chandni Ghatak, “Artificial Intelligence and Intellectual
Property Law”, Christ University Law Journal Vol.7, No.1, (2018).
Tatiana Zaplatina, “the Legal Issues of the Artificial Intelligence, Intellectual
Property Rights and Data Protection”, Kutafi n University Law Review Vol 6
Issue 1 (2019), p. 45
Tomoko Miyamoto. Patentability of Computer Software and Business Methods.
PowerPoint Presentation: delivered at WIPO-MOST Intermediate Training
Courses on Practical Intellectual Property Issues in Business, November
10–14, (2003).
128 B. A. Riswandi
USPTO. Inventing AI: Tracing the Diffusion of Artificial Intelligence with

U.S. Patents. U.S. Patent and Trademark Office-Office of the Chief
Economist IP Data Highlights, 2020.
World Intellectual Property Organization. WIPO Technology Trends 2019:
Artificial Intelligence, World Intellectual Property Organization. Geneva-
Switzerland: WIPO, 2019.
World Intellectual Property Organization. WIPO Conversation on Intellectual
Porpoerty (IP) and Artificial Intelligence, Document Summary by WIPO
Secretariat, January, 2021.
Regulation
Indonesia Law No.13 of 2016 concerning Patent

Patent Act 35 U.S. Code
特許法 Tokkyohō
Trade-Related Aspects of Intellectual Property Rights (TRIPs) Agreement.
Care Robots for the Elderly: Legal,
Ethical Considerations and Regulatory
Strategies
Hui Yun Chan and Anantharaman Muralidharan
1 Introduction
1.1 An Ageing Population
It is projected that by 2030 there will be around 1.4 billion people aged
60 years and older and doubling to 2.1 billion by 2050 (WHO 2022).
High income countries such as Japan and the US have a substantial pro-
portion of people who are 60 years and older. As people continue to live
longer, people who are 80 years and older will continue to grow to
426 million by 2050 (WHO 2022). Older people live in various types of
household settings, ranging from living alone or in communities to those
who live in nursing or care homes requiring some form of support in
their daily lives. Some of them could be recovering from illness or in
H. Y. Chan (*) • A. Muralidharan

Centre for Biomedical Ethics, Yong Loo Lin School of Medicine, National
University of Singapore, Clinical Research Centre, Singapore, Singapore
e-mail: c.hy@nus.edu.sg; murali@nus.edu.sg
https://doi.org/10.1007/978-3-031-42576-9_6
130 H. Y. Chan and A. Muralidharan
various stages of rehabilitation, while some maintain an active post-

retirement life.
An ageing population gives rise to a range of challenges (Van Zaalen
et al. 2018). Caring for the elderly requires increasing resources, ranging
from caregivers to hospital and community resources, with implications
for governments in creating innovative ways to provide continuous care
for the ageing population. (Garner et al. 2016; Fulbright 2021; Wilks
et al. 2005, 37).
Care robots to the rescue? Prevalence, objections, and concerns regard-
ing care robots
While some of the caregiving labours are outsourced to home-based
domestic helpers or live-in nurses and carers, some countries have turned
to care robots to assist with care delivery in various capacities to enable
the elderly to continue living in their familiar environment and being
close to relatives. This development, combined with advancements in
artificial intelligence, software engineering and robotics have contributed
to progress in developing AI-integrated care robots, pushing the frontiers
in care delivery. The ‘Elderly Care Giver’ concept, for instance, is one of
the many longstanding efforts by robotics developers that is geared
towards creating multitasking care robots intended as personal assistants
for the elderly (van Aerschot and Parviainen 2020). Another type of care
robot known as artificial companions are robots with certain levels of
intelligence and social skills sufficient to create and sustain long-term
relationships with their users (Lim 2012). Their functions range from
fulfilling human desires for social connections to providing information
and communication, as well as entertainment (Floridi 2008, 652, 653).
Other types of robots include those intended for patient rehabilitation
and exoskeleton (Tan and Taeihagh 2020). For example, in Singapore
various initiatives under the Singapore Health Assistive and Robotics
Programme were implemented to integrate care robots with distinct
functionalities into different aspects of care provision, ranging from
robotic nursing assistant, care assistant and rehabilitation enabling robot
to AI-driven robotics for last-last mile delivery (NHIC 2022).
Within Europe and North America, innovations in care robots include
those that are designed specifically to fulfil certain tasks. For example,
CALONIS, a type of conversational artificial companion is developed for
Care Robots for the Elderly: Legal, Ethical Considerations… 131
the purpose of engaging and interacting with former soldiers who suf-
fered from brain trauma with cognitive injuries (Wilks et al. 2015). These
care robots are programmed to remind the users about appointments,
and other activities in their daily lives such as brushing their teeth.
CALONIS was modelled after a European Union- funded Senior
Companion project which was devised to interact with older people,
where these care robots learned about the lives of the elderly via photos,
and conversations (Wilks et al. 2015, 258). Other examples include vir-
tual carers for the elderly that aids their daily living and interacts with
them (Garner et al. 2016) and LOIS, an elderly companion intended to
assist the elderly in managing their needs and monitor their wellbeing on
a long-term basis (Fulbright 2021, 403).
Some scholars observed that despite continuous efforts and scientific
advancements, the production and use of care robots remain minimal
(Hoppe et al. 2022; Fulbright 2021, 404). These could be attributed to
manufacturing obstacles that are largely focused on producing robots
that carry out simple tasks and research and development of care robots
on a smaller scale for specific research purposes that are limited to labora-
tory settings (Van Aerschot and Parviainen 2020). Gaps in regulation,
financing, and data security are among some of the factors influencing
low level of trust in using care robots (Hoppe et al. 2022). A change in
how care robots is used in the world could potentially aid a wider use of
care robots. Van Aerschot and Parviainen (2020, 253) proposed that their
use should be framed within the entire care structure encompassing an
interdependent system of emotional, social and practical contexts of care-
giving. In enabling a widespread use of care robots, they recommended a
realistic approach to research that considers market drivers and needs
(ibid, 254). This approach could be valuable to inform how regulators
shape governance directions for care robots.
What about user experience in interacting with care robots? Some
studies reported positive perceptions towards care robots among caregiv-
ers in elderly homes (Rantanen et al. 2020; Niemela and Melkas 2019).
Adequate training to build competency in interacting with care robots,
incentives and usability are said to encourage the adoption of care robots
by end users (Frennert et al. 2021). A study exploring users’ experience
with care robots however revealed the need to be cautious of potential
differences between robots that are tested and operated in lab environ-
ments and robots that are integrated in home settings (von der Putten
et al. 2011, 333, 334). Such findings may affect how users perceive care
robots in their home settings, how these interactions change the users
and the inclination to build relationships with them in spite of their arti-
ficial nature. This seems to support a greater level of comprehensive
awareness of underlying beliefs that humans and robots have in creating
care robots through an observation of how they interact in the real world,
rather than in labs (Heylen et al. 2011). More recent research from
Sweden reported stakeholders’ doubts about the extent of readiness on
the part of their national governance framework within the social and
systemic contexts in introducing care robots as part of elder care
(Johansson-Pajala and Gustafsson 2022). Further studies exploring per-
ceptions of politicians, insurance organisations and the media in Finland,
Sweden and Germany similarly referenced the significance of regulations
in supporting the use of care robots in areas such as employment, avail-
ability of care robots and financing to developers (Hoppe et al. 2022).
There are, however, some resistance to the use of care robots. Sharkey
and Sharkey (2012) raised ethical concerns regarding care robots ranging
from reduction in human contact, vulnerabilities arising from loss of
control and objectification, loss of privacy, personal liberty, and decep-
tion. The inability of care robots to meet the emotional needs of the
elderly due to a lack of authentic social interaction and the capability to
establish affective relations (Sparrow and Sparrow 2006) are further
objections to care robots. It is claimed that robots are incapable of respect-
ing or recognising users and such right to be respected would be under-
mined should they be allowed (Sparrow 2016, 446). It can be said that a
common objection to affective technologies such as social robots and
those with close human interactions such as care robots is the existence of
associated risks with their use such as vulnerability, interdependency,
responsiveness and responsibility, which gives rise to wider implications
on caregiving relationships among human beings.
In response to criticisms against care robots, some scholars emphasised
the importance of empathy to enable the creation of relationships between
care robots and their human users (Leiten et al. 2013; Coeckelbergh
2010). Chan (2021, 638) similarly advocated for care ethics in the design
of care robots for its value in fostering trust and its suitability both in the
caregiving settings as well as alignment with regulatory guidelines aimed
at promoting the safety and security of users, respect for human rights,
dignity and privacy and accountability of developers. These recommen-
dations could be seen as attempts at defending against claims that robots
are incapable of caring.
These concerns remain as long as robotics continue to develop. Care
robots have the potential to affect people’s health and safety, as they are
placed within the private home environments of the elderly who are
likely to be vulnerable. Their daily lives, relationships, object of depen-
dency and experiences could be transformed, thus raising concerns about
undesirable effects these robots might have upon the users (McTear et al.
2016, 296, 297). Some research suggested that the aims of robotics
should be aligned with human rights principles and relevant ethical val-
ues of privacy and autonomy (Vargas et al. 2011, 330). Practical and legal
questions such as potential changes to liability regimes where harms
occur from the use of care robots are set to become more important as
innovations continue (Wilks et al. 2005) as well as questions regarding
the legal personality of robots and attribution of responsibilities in the
event of harm (Begishev et al. 2021). Whilst their use may be regarded as
beneficial from a scientific perspective, important ethical and legal issues
need to be considered.
1.2 Ethical Complexities
Vandemeulebrouckea, de Casterleb and Gastmans (2018) in their sys-

tematic review of ethical arguments regarding care robots highlighted a
range of ethical arguments that are primarily influenced by ethical discus-
sion within aged care settings. The authors proposed the application of
‘democratic spaces’ that involves all stakeholders in aged care in these
debates to better assess their use and reflect on their implications to the
users (24). This approach enables a broader consideration of other impor-
tant influences affecting the relationship between care robots and the
elderly, such as issues of reliance, autonomy and responsibility. Two rele-
vant ethical aspects are discussed below.
1.3 Deception and Manipulation
An AI-based robot that is designed to take care of the elderly is intended

to perform several tasks. These include administering medicine and treat-
ments, assisting them with daily living tasks and having conversations
with them. Fulfilling these tasks requires trust. After all, if the elderly
person did not trust the robot they would not cooperate with the robot.
For instance, in order for the robot to fulfil its function of administering
medicines, the elderly person needs to be able to trust that the mixture of
pills he or she is receiving contain the correct medicines at the correct
dosages. Distrust here would result in treatment non-compliance and
thus the failure of the robot in its task.
It seems natural at this point to find ways to increase the trust that the
elderly invests in these robots by various means. Of particular ethical
concern, are methods that can be regarded as manipulative or deceptive.
In this section, we examine attempts to boost trust in AI by making the
robot more humanoid or shaped like cute animals (Siau and Wang 2018)
and why such attempts would be considered manipulative. We might
initially think that making these AIs look humanoid or similar to cute
animals attempts to deceive people into thinking that it is a person or
cute animal. However, merely being humanoid or dog-shaped does not
mean that one is attempting to deceive people into thinking they are
humans or dogs. One would have to go much further in masking its
obvious artifice in order to even count as trying to deceive.
While most of us would not find these AIs deceptive, the elderly who
may be at some stage of dementia might be deceived. Deception, after all,
is ordinarily wrong because it attempts to use someone’s capacity to rea-
son for our own ends. This thus fails to respect their capacity to reason.
However, for the elderly with dementia, their capacity to reason is
impaired. Since respect, in the sense that we are interested in, is a matter
of recognition (Darwall 1977), disrespect must be a matter of misrecog-
nition. That is, we disrespect someone only if we treat them as lacking the
capacity to reason when they actually possess such capacity. However,
here, we are arguably not misrecognising the elderly’s capacity to reason,
but simply acknowledging that they lack this capacity to an adequate
degree and hence should be treated accordingly. Moreover, we are only

getting them to trust the robots so that they will willingly take their med-
icines. Thus, arguably, we are not using them for our own ends, but only
getting them to do what they would want to do if their cognitive capaci-
ties were intact. While this refinement of the initial argument (viz decep-
tion) does not succeed, it and the response to it point to why such ways
of increasing trust are objectionable. Firstly, the refinement makes it clear
that the “increase trust by deception” assumption is mistaken. If trust was
improved via deception, trust would increase only among those who are
susceptible to believing that the robots were human like those suffering
from dementia. Yet, this is not what was observed in Siau and Wang’s
study where levels of trust increased for all groups.
Secondly, trust in AI is being generated for reasons that are distinct
from the trustworthiness of the AI in question. People are not increasing
their trust because they learn something or come to believe something
that, if true, would give them more reason to trust AI. Instead, they are
being led to trust AI more because it is packaged in certain ways. Looking
at this through the lens of nudges can be illuminating. Nudges exploit the
ways in which our choices are packaged in order to influence our choices,
for good or for ill. To be clear, not all nudges are manipulative in any
interesting sense. “Bolding” allergens in an ingredients list, for instance,
is a nudge that influences choices by highlighting information that we
might be otherwise prone to overlook. Just as not all nudges are manipu-
lative, we might, following Sunstein (2015), note that not every nudge
that we might call manipulative is necessarily morally objectionable. For
instance, consider an innocuous handphone advertisement which uses
catchy music and bright colours to frame the product. While this adver-
tisement is manipulative in that it draws on non-rational considerations
to influence choice, it is not objectionably manipulative. Objectionably
manipulative nudges are paternalistic, do not provide subjects with salient
information, and interfere with deliberation. In this vein, what is wrong
with increasing trust in AI by packaging it in humanoid or cute shapes is
that it is objectionably manipulative because it seeks to increase trust for
no good reason (Sunstein 2015).
Defenders of these types of care robots might object at this point that
they are not trying to induce trust by presenting irrelevant considerations.
Instead, they would be attempting to negate irrelevant sources of distrust.

After all, even when told that a human and AI are equally reliable at a
task, people are inclined to trust the human over the AI (Juravle et al.
2020). The objection here is that this distrust of AI is unjustified and
must be engendered by some irrelevant factor. The purpose of shaping AI
assistants to more humanoid or cute forms is to remove an inadvertent
factor which would influence patients through irrelevant factors. The
thought is that instead of illicitly increasing patient trust, they are only
preventing the illicit reduction of patient trust. It does not follow from
the fact that people trust persons more than AIs that people illicitly
decrease their trust in AI. It could be that people illicitly trust people too
much or that the gap in trust is justified. In fact, we may have more rea-
son, everything else being equal, to trust people than similarly reliable
AIs. After all, while we have reasons of reliability to trust both, we also
have moral reasons to trust people, but not AI.
In support of such moral reasons, consider the often heard injunction
to give people the benefit of the doubt. For instance, it used to be com-
mon, perhaps in kinder times, for members of the opposition to give the
winning candidate the benefit of the doubt and also a chance to lead the
country (Jones 2004). Such giving of the benefit of the doubt trusting in
the face of some evidence of unreliability. This suggests that we think
there are some moral reasons to trust persons even when considerations
of reliability would suggest not trusting. If there are such moral reasons,
we can see why we can permissibly trust persons more than equally reli-
able AIs. After all, here, the moral reasons stack up. One might object
that these moral reasons are reasons of the wrong kind. On this view,
trust must be apportioned to reliability. However, this is far from clear.
Suppose, for instance, that Alf, a recovering alcoholic, is at an office party
and a colleague spills wine on his shirt. When he returns home, his wife,
Betty, believes that he has fallen off the wagon despite his protestations
otherwise. It seems that Alf can legitimately expect Betty to trust him
even when she has evidence (his past alcoholism and the stained shirt)
that he is not to be trusted on this matter (Holton 1994). Even if Alf did
have a sip of wine, we might think that there is still something problem-
atic about Betty’s failure to trust Alf. If Betty has reason to trust Alf, it
cannot be based on any sort of evidential or alethic considerations. After
all, if they were evidential or alethic, we would think that there is nothing
wrong with Betty not trusting Alf. As such, the reasons must be of a
moral kind. For instance we might think that Betty owes it to Alf in vir-
tue of him being her husband. Or we might think it is good for Betty to
trust Alf so that Alf can learn to develop responsibility or trustworthiness
with regards to his drinking problem. These moral reasons make trust apt
in these circumstances despite Alf ’s lack of trustworthiness. This means
that moral reasons cannot be the wrong kind of reasons to trust. It fol-
lows then that there are at least some cases where we may permissibly
trust a person more than an equally reliable AI.
Given that such a trust-gap can be justified, it is not altogether clear
that increasing trust by shaping AI assistants in humanoid or cute forms
is merely a matter of removing illicit sources of distrust. Instead, it can
plausibly, be described as an illicit attempt to engender trust. After all,
being humanoid or cute is not any sort of reason at all to trust some-
thing however being a person is.
1.4 The ‘Zombie Relationship’ Problem
One type of use these AI assistants or care robots may be put to is engag-
ing or at least attempting to engage the elderly in conversations. Let us
call these types of assistants, conversation bots. In using conversation
bots, attention needs to be paid to the value of such assistance. Certainly,
one aspect of this value is instrumental: There are a number of health
problems and risks associated with social isolation (National Academy of
Sciences, Engineering and Medicine 2020). Conversation bots are sup-
posed to be useful in mitigating some of these risks. While the health
risks associated with social isolation are serious, mitigating health risks is
incidental to the practice of having conversations and social relationships.
Instead, the key point of having conversations and social relationships is
to connect with a real person on the other side. This aspect matters for
two reasons. Firstly, it matters because a large part of the value of having
social relationships is grounded in having a relationship with an actual
person. Secondly, it matters because people are not likely to engage in
conversations purely for health benefits. Instead they engage in
conversations in order to communicate with other persons. As we will

argue, the health benefits of social contact plausibly arise from the engage-
ment (real or imagined) with other minds. This means that conversa-
tional AI may be of dubious value and may even be harmful if used to
replace human contact.
Let us first establish the value of there being a person on the other side
of the conversation or relationship. To do this, we shall make a small
digression: In order to argue against the idea that mental facts were reduc-
ible physical facts, David Chalmers (1996) introduces us to the zombie
argument. He argues that since we can imagine beings who are physically
identical to us, down to the last atom, but who lack conscious experience,
mental facts are not reducible to physical facts. Chalmers calls such beings
zombies. We are not concerned with whether Chalmers’ argument is suc-
cessful. Instead, we use the concept of zombies for a different purpose.
Imagine that zombies, i.e., beings which look and behave human, but
which have no conscious experiences, exist among us. Since they out-
wardly look and behave human, there is no easy way to distinguish zom-
bies from humans. Consider a situation where one person, Alf, believes
himself to be in a relationship with Betty who, unbeknownst to him, is a
zombie. Intuitively Alf ’s relationship with Betty is a sham. In order for
the relationship to be real, Alf and Betty need to love each other. However,
Betty cannot feel love since she has no conscious experience. Hence, she
cannot love Alf. Likewise, Betty’s statement that she enjoyed their date is
also false since Betty is incapable of enjoyment. After all, while Betty can
react and produce appropriate pain and pleasure responses to the respec-
tive stimuli, she feels neither pain nor pleasure.
Hence, if Alf were to find out that Betty is a zombie, Alf would have
no reason to ask Betty how her day was or ask for her opinion on an issue.
He has no reason to keep his promises to her as there is no one to whom
he owes the keeping of said promises. He has no reason to be tactful or to
compromise on any issue. After all, she has no feelings to be hurt or pref-
erences or values that need accommodation. If this is true in Alf ’s case, it
would be even more so for the elderly who interact with conversation
bots. After all, while Alf could choose to continue doing all the usual
relationship things because Betty cooks for him, pays half the rent and
provides physical intimacy, the elderly have no such pecuniary
considerations with respect to a conversation bot. At this point, one

might object that conversation bots are no worse than human nurses who
are paid to care for (in a practical sense) the elderly but who may not care
(in an emotional sense) about their charges (Lancaster 2019). The thought
here is that the apparently emotionally rich relationship that an elderly
person thinks they have with such a nurse is just as much a sham as it
would be if the nurse turned out to be a cleverly disguised robot or a
zombie.This objection is mistaken because while the specific emotional
relationship the patient believes she has with her nurse is a sham, there
still exists a transactional relationship with the nurse. And while such a
purely transactional relationship may not be as valuable as one which is
more emotionally laden, it is still better than having no kind of relation-
ship at all as would be the case with a care bot. The nurse still has some
reason even if only instrumental reason to care about her patient’s wellbe-
ing. Moreover, with human nurses, it is possible and perhaps quite com-
mon for a more than merely transactional relationship to exist between a
nurse and her charge. By contrast, there is no such possibility for a con-
versation bot or a zombie. In the latter two cases there will necessarily be
no transactional relationship let alone an emotionally laden one with the
patient. This last point is crucial. Conversation bots do not have minds.
Even if we thought that the health benefits would still be some reason to
employ such AI assistants, the elderly will not enjoy these benefits if they
refuse to engage in conversations with the bots. To see why this might be
the case, consider how social isolation can harm people and correspond-
ingly why social interaction staves off those harms. There are two broad
classes of causal mediators we might consider: Subjective and objective
(National Academy of Sciences, Engineering and Medicine, 2020). A
subjective source of harm would be the feeling of loneliness that people
experience from knowing that they have no one to talk to, nobody to rely
on etc. Insofar as the elderly know that they are talking to robots, it is
hard to see how they could not feel lonely since they do not have any
relationship with any person. That is to say providing a robot does not
seem like it could relieve feelings of loneliness where the patient in ques-
tion is fully aware that the entity they are conversing with is not a per-
son. Even if it could relieve such feelings, it is not clear that this would be
a genuine improvement in patient well-being. After all, feeling lonely is,
in fact, the apt response to involuntary solitude. This is similar to how

fear is the apt response to something being dangerous and grief is the apt
response to the recent loss of a loved one. The mere fact that these are
negative emotions does not mean that it is desirable, all things consid-
ered, to relieve a person of those feelings without changing the underly-
ing reality which made those feelings apt. If so, even if care-bots could
relieve feelings of loneliness among the elderly who are fully aware that it
is a robot, this would be perverse.This brings us to the objective causes of
harm. Broadly speaking objective causes of harm are those which are
caused by the lack of relations with actual persons. It is worth paying
attention to the way in which social interaction helps stave off cognitive
decline in the elderly. It is well established that having cognitively com-
plex lifestyles can help stave off cognitive decline among the elderly
(Festini 2022). Consider now, how we engage in conversations with oth-
ers. When we talk to others, we try to watch what we say in order to avoid
offending the other person. We try to give reasons in order to convince
the other person of our own position. We pay attention to the other per-
son’s verbal and non-verbal cues in order to understand what the person
is trying to say and engage with someone who has their own thoughts
and feelings. These tasks can be cognitively demanding. Small wonder,
then, that social isolation contributes to cognitive decline. Social isola-
tion deprives people of a significant source of cognitive stimulus and this
can hasten cognitive decline.Where the elderly know that these are bots
and hence lack minds, they should have no reason to converse with the
bots in the way they converse with persons. We perform the cognitively
demanding tasks during conversations because our conversational part-
ners are persons and in general we have respect our interlocutors by not
needlessly offending them, by giving them reasons for our belief and by
trying to understand what they are saying and where they are coming
from. In the case where the elderly know that their “interlocutor” is not
a person, there is no reason to respect the robot and hence there is no
reason to watch what they say, give reasons or pay attention to verbal and
non-verbal cues. Without exercising these conversational skills, even if
the elderly person does not outright ignore the care-bot, any “conversa-
tion” between them is unlikely to contribute to a cognitively complex
lifestyle which would aid in staving off cognitive decline. This leaves us
with the option of getting the elderly to engage with these AI assistants
by deceiving them into thinking that they are conversing with humans.
However, as noted, this would be explicitly deceptive and hence imper-
missible for that reason. If, on the other hand, we remained within ethi-
cal boundaries, it is not clear what benefit such conversation bots would
have since they are unlikely to be used. Moreover, we can see how replac-
ing nurses with conversation bots may even lead to harm. Whereas a
human nurse could genuinely engage the conversational skills of their
charges, it is not clear that conversation bots could do the same to the
requisite degree outside of a laboratory setting.
The ethical considerations discussed so far examined the relationship
complexities between how care robots interact with the elderly that affect
how they ought to be designed and developed. An elderly person is not
automatically vulnerable and lacking in discerning abilities, however, to
prevent deception and manipulation for the elderly who are more likely
to be vulnerable (e.g.: those suffering from dementia), care robots could
be designed to be less human-like as much as possible to ameliorate the
deception and avoid undue attachment to these robots. Further, from a
safety perspective, the more human-like the care robot is, the more agile
it has to be, which will create hazards when placed in the elderly’s home
environment. The priority then should be to ensure that these robots are
functional, stable, and safe to use rather than aesthetically pleasing.
2 Governance Framework for Care Robots

2.1 Legal Complexities and Considerations
Human robot collaborations intended to improve the quality of life of

the elderly with daily interactional functionalities created a slew of legal,
ethical, practical and governance concerns. Existing regulatory landscapes
remain fuzzy with fragmentary guidance, and ad-hoc standards to address
emerging risks. The ideal regulatory regime may yet materialise, and
indeed may not be feasible, however the existence of adequate regulatory
oversight is crucial given the aim of governance is not to stifle innovations
but to ensure that care robots that are marketed and integrated into the
lives of elderly are safe for their use and that the elderly users are secure
when interacting with care robots. Regulatory bodies continue to face
questions such as: how should the legal framework accommodate this
new technology? Are existing laws or guidance sufficient to address this
new application? These and many related questions will continue to
remain important while newer care robots are developed, tested, and
marketed to the relevant sectors.
One of the complexities with designing regulatory options is address-
ing the variety of care robots’ functionalities, purposes, software, and
hardware compositions. The sophistication and development of care
robots meant that it could cross several regulatory scopes, which may give
rise to duplicity or in some instances fall outside of the regulatory scope.
For example, what types of law should apply to a care robot that assist
with the daily routine and can interact with an elderly who is rehabilitat-
ing from stroke or managing a progressive illness such as Alzheimer’s or
dementia? Is the care robot treated as a medical device, or assistive tech-
nology capable of providing information useful for medical diagnoses
and planning future treatment? When care robots learn more informa-
tion about the elderly that they cared for and store or transfer such data,
it becomes clear that such care robots cannot be precisely classified under
laws that govern consumer use of medical devices solely. Further, con-
cerns about security of collected data, privacy of the elderly and families
and how these data are managed, maintained and used by healthcare
providers require additional consideration of applicable laws regarding
data protection and security.
Another relevant consideration is the potential modification to con-
ventional doctor- patient relationships as care robots have seemingly
become the intermediary between the elderly patient and healthcare pro-
viders, and sometimes an integral part of the treatment or rehabilitation
process. Rules relating to doctor-patient confidentiality may have to be
extended to these care robots or revised to accommodate the change in
this relationship. Other important concerns such as trust, and disclosure
of information are similarly affected where care robots enter these health-
care settings (Fosch-Villaronga 2020, 176). It will be seen that Article 9
of GDPR relating to processing of health-related data can become
applicable as providers justify the exceptional ground upon which these

data processing can occur.
The following sections highlight key concerns that regulators face in
designing appropriate governance responses.
2.2 User Health, Safety and Wellbeing
The extent of interactions between care robots and the elderly users affect
the latter’s physical safety and emotional wellbeing. User health, safety
and wellbeing implicate legal considerations (Nambu 2016, 484). The
physical appearance of care robots may not mimic an actual human
being, however long periods of interactivities may create ethical issues
that are hidden at first instance. The issue of deceptiveness of care robots,
as discussed above and the inherent vulnerability of elderly who are
dependent on these robots as part of their ‘normal’ functioning in daily
routines augment their susceptibility to risks. There is a need to ensure
that care robots are safe to use within elderly populations where there is a
higher risk and degree of dependency on these robots compared to other
sectors. Doubts about the most appropriate classification for care robots –
whether they are medical devices or general lifestyle gadgets remain prob-
lematic for regulatory purposes as different intended functions attract
different liability considerations (Fosch-Villaronga and Mahler 2021).
Developers and creators need certainty in comprehending their obliga-
tions and protections under applicable standards for safety requirements
to enable legal compliance to ensure that users are truly protected (Fosch-
Villaronga 2016, 35, 42). Strengthening legislative protections and pro-
moting transparency in how these care robots are used and the effect on
the elderly are potential safeguards to ensure the safety and wellbeing of
elderly users.
2.3 Data Security and Privacy Protection
Data security, cybersecurity and privacy protection are forefront issues in

a big data environment and are particularly crucial in the deployment of
care robots (Pierce 2020; Fosch-Villaronga 2020, 173; Fosch-Villaronga

and Mahler 2021; Tan and Taeihagh 2020, 15). Care robots that are
designed to interact with their users are capable of collecting and storing
user information which can be subsequently transferred or communi-
cated to healthcare providers, cloud service providers or manufacturers
for further processing and use (Holder et al. 2016, 391). It is this capabil-
ity that gives rise to data security threat as the collected information may
include highly personal and potentially sensitive information such as
behavioural, emotional and physiological information, that are liable to
be protected from unauthorised disclosure. Data security and privacy
risks range from hackers gaining access and creating false profiles to
deceive users or their relatives to potential breaches in security protocols
and information theft that render users vulnerable to scams (Garner
et al. 2016).
It becomes significant that users are made aware of these risks and
providers to be responsible in ensuring these risks are addressed or mini-
mised. The elderly should be informed about how their data is collected
and used by care robots and caregivers who need to provide treatment or
rehabilitation assistance. Nonetheless, the lack of specific guidance about
ways to manage security risks despite the existence of cybersecurity certi-
fication framework at the European Union level meant that users con-
tinue to be at risk as designers and developers are uncertain about the
level of minimum standards that they ought to comply with to ensure
user safety (Fosch-Villaronga and Mahler 2021). The implications that
result from this approach include inadequate IT, software updates and
security protocols in places where care robots are used, such as hospitals,
care homes and caregivers’ settings. New cybersecurity rules for new
products such as care robots may be valuable, meanwhile the gap can be
addressed by implementing a Manufacturer Disclosure Statement for
Medical Device Security (MDS2) form to manage security concerns
involving authentication, software upgrades, malware detection (Fosch-
Villaronga and Mahler 2021).
Reliance on existing data protection rules such as the GDPR may help
determine appropriate roles and responsibilities of data controllers and
those who have access to these potentially sensitive information. For
example, it is necessary to identify who is the data controller or processor
to enable a clear demarcation of obligations in securing and protecting

these collected data. Consistent with these obligations, Article 8 on the
right to private life is applicable as vast amount of personal data is pro-
cessed. Prior to collecting data, developers and manufacturers should
consider implementing privacy impact assessments, privacy by design,
security by design and privacy by default to help address some of the risks
arising from processing large amounts of personal data (Holder et al.
2016, 394). The approaches outlined so far support active human inter-
ventions in all stages of these processing to properly ensure the lawful
collection and processing of these personal data.
2.4 Balancing Innovations, Market and Protecting

User Safety
The struggle for policy makers is not new as some studies from Australia
and New Zealand have shown that there is limited governmental capacity
in steering policy developments to provide oversight for robot technolo-
gies arising from limited capabilities in comprehending important reper-
cussions from adopting robotics technologies within a complex system of
markets, regulation and technological skills. (Dickinson et al. 2022).
Japan grappled with the best way to regulate care robots given that there
is an absence of clear definition for care robots within its legal framework
(Nambu 2016, 484) as well as technology outpacing regulatory scopes
(Iizuka and Ikeda 2021). It is recognised that the difficulty in defining
care robots may well stem from concerns that any definitions could
potentially create unintended obstacles for innovations and commerciali-
sation strategies (Nambu 2016). The lack of established standards creates
difficulties in properly identifying appropriate liability regimes when
harms occur from using care robots. In attempts to balance these two
interests, Japan introduced its New Growth Strategy in 2010 aimed at
ensuring user safety in their interactions with care robots through manag-
ing user feedback and contributing to safety standards (ibid, 494, 495).
Prior to this Strategy, Japan followed the ISO standards concerning safety
guidelines for next generation robots in 2007 (ibid, 488). The publica-
tion of Japan’s Robot Strategy in 2015 provided further regulatory
impetus to support the safe and extensive use robots in everyday life while
advancing technological innovations in robotics (ibid, 496). Other sug-
gestions include a greater involvement of various regulatory institutions
and creation of new roles for governments in regulating care robots
(Iizuka and Ikeda 2021) and more cross departmental collaborations in
developing policy response for care robots (Chou et al. 2019). ISO stan-
dards appeared to be a favourable option in regulating safety standards
for care robots as it has the advantage of enabling various stakeholders to
traverse governance processes from an international source point where
manufacturers and developers ought to comply with (Iizuka and Ikeda
2021). The situation regarding regulatory difficulties in Japan is not too
dissimilar from Taiwan where there is a lack of specific care robot policies
and an absence of policy coordination and involvements of various gov-
ernmental departments (Chou et al. 2019).
The balancing approach is influenced by the purpose envisioned by the
regulators when devising regulatory guidance. Reference can be made to
the distinctions between the EU certification approach and the US Food
and Drug Administration (FDA) policy in regulating medical devices.
The EU system (such as a-minimum standards approach in ISO
13482:2014) favours promoting scientific advancements while the FDA
is inclined towards protecting consumers from harm, resulting in the dif-
ferentiated assessment processes (Fosch-Villaronga 2020). One such
aspect is in the classifications of care robots, which has remained unset-
tled. For example, the FDA categorised medical robots as class 2 medical
devices (considered as medium risk), such as surgical robotic systems or
robots that function to improve the quality of life of patients such as
Paro, a robot to help patients with dementia and Alzheimer’s. However,
the increasing sophistication of intelligence in care robots may add fur-
ther complexities in the assessment of risks concerning its autonomous
nature. This approach is due to augmented developments in robotics,
which affect pre-market approval stage costs in testing, resulting in regu-
lators shifting responsibilities to manufacturers in demonstrating safety
equivalence with devices already on the market and self-certification
(Fosch-Villaronga 2020, 3). Those that fall within a higher risk category
will be certified by an external body. This approach however does not
resolve the uncertainties in terms of compliance with medical device laws.
Striking an appropriate balance remains complicated as other consid-

erations are influential factors. For example, socio-cultural acceptance to
risks from care robots vary across different types of populations, conse-
quently affecting regulatory scope and purpose. In the future, where spe-
cific laws are yet to be created, there is an opportunity to include users’
and society’s preferences about care robots where they are sought to be
implemented (Tan and Taeihagh 2020, 15). An approach that prioritises
stakeholders’ views is more likely to result in an approach that effectively
protects users from harm when interacting with new technologies.
Research has alluded to the importance of attending to the abilities and
needs of users of care robots to ensure designers incorporate safety con-
siderations in designing and implementing these robots (Caleb-Solly
et al. 2021). As such, it is crucial to promote evidence gathering and shar-
ing to support stronger regulatory justifications through transparency,
advancement of knowledge and cross-agency collaborations combined
with effective coordination between national and international regulators
in developing regulatory responses (Iizuka and Ikeda 2021). Regulations
that did not recognise the mutuality of design, human needs and regula-
tory aims risk being ineffective, consequently, regulations should be
devised in a way that promotes a cyclical nature of communication
between developers of new technologies and regulators to improve the
regulatory effect towards producing feasible laws (Fosch-Villaronga and
Ozcan 2020, 960, 968).
Responsive regulations are important to address some of the complexi-
ties in governing these technologies through appropriate governmental
stewardship (Dickinson et al. 2022). One of the suggestions as to how
care robots can be regulated to balance both interests is for innovations to
influence regulation where governments offer responsive governance
rather than the traditional top-down approach of regulations affecting
innovations (Iizuka and Ikeda 2021, 102335). However, this might be
perceived as being closely influenced by market considerations that risk
user safety concerns. This approach can be balanced if supported by active
stakeholder oversight. Whilst it is important to support innovations, gov-
ernments have the responsibility to protect the public through regula-
tions that steer innovations towards responsible and publicly appropriate
ends. It is preferable that governments play an active regulatory role to
govern robotics applications (Palmerini et al. 2016). Dynamic regulatory

options that cater to various risks that arise at distinct levels of techno-
logical developments may be a viable and feasible option that does not
curb innovations yet enable proactive oversight of any issues that arise.
2.5 Regulatory Strategies and Options
The range of regulatory guidance currently available provides further

considerations for formulating, revising, or establishing rules that are bet-
ter tailored to research, development, marketing and use of care robots
for the elderly population. Existing guidelines that are applicable to care
robots include the IEEE Global Initiative on Ethics of Autonomous and
Intelligent Systems, Ethically Aligned Design: A Vision for Prioritizing
Human Well-being with Autonomous and Intelligent Systems which
focused on general principles of prioritising human rights and human
wellbeing in designing and using care robots, developers’ accountability
and transparency and minimising risks of misuse (IEEE 2019). These
principles share some similarities with the British Standards Institute
Ethical Design of Robots encompassing risk management, safe design
and information for the use of care robots.
The European Commission Robolaw project focused on the legal and
ethical concerns regarding consumer health and safety, liability, intellec-
tual property rights and privacy and data protection and protection of
basic rights (Fosch-Villaronga 2020, 91, 95), covering the breadth of
essential rights ranging from respecting basic tenets of rights, autonomy,
privacy, justice and social connectedness. While they are generally appli-
cable to any types of robotics applications, these widely accepted values
are relevant to care robots. The Responsible Robotics organisation simi-
larly developed seven principles for responsible robotics that are broadly
applicable to care robots: security, safety, privacy, fairness, sustainability,
accountability and transparency (FRR 2022). The UN Convention on
the Rights of Persons with Disabilities which emphasises respect and sup-
port for independent living is applicable for the elderly who have disabili-
ties and may benefit from using care robots. Compliance with established
laws and accepted tenets of privacy and related rights remain important
in the design process as they affect user safety, especially for vulnerable
populations (Chan 2021, 638).
Besides regulatory standards, a more specific-oriented approach in the
form of a robot impact assessment for care robots could potentially ame-
liorate emerging safety risks that arise such as user privacy, data security
and liability issues (Fosch-Villaronga 2020, 5). The assessment is under-
pinned by six principles: impact explanation, comprehensive safety:
physical, cognitive safety requirements, consumer robotics: health, con-
sumer protection and environmental protection, liability: current and
prospective responsibility, accountability and liability, privacy: privacy
and data protection and dignity and broader implications: independence
and autonomy, dignity, ethics, and justice (ibid, 97). These principles are
consistent with existing international guidance and serve to support their
operationalisation. As highlighted earlier, care robots that are used in dif-
ferent settings and contexts create distinct risks and concerns, as such a
robot impact assessment is valuable to identify these distinct risks to
enable a timelier response. In addition to such functional impact assess-
ment, a regulatory sandbox approach trialled in Singapore is useful to
advance innovations but also identify risks that may be hidden when
robotics applications are tested in laboratory settings (Tan and Taeihagh
2020). This sandbox approach enables the creation of robotics testbeds
and pilots in health clusters across the country, with feedback from users
and providers to improve the applications and address concerns that arise.
A broader, overarching type of regulatory approach that strives to
embed adaptive elements is an iterative regulatory process proposed by
Fosch-Villaronga and Heldeweg (2018). They identified a coordination
gap between regulators and robot developers that resulted in disjointed
regulatory outcomes that neither effectively support innovations nor pro-
tect users of care robots. Pursuant to their iterative regulatory approach,
they proposed using forecasts from robot impact assessment for ethico-
legal evaluation and actual results of legislative assessment to revise, mod-
ify and update the laws. These are intended to create an active
evidence-based policy for emerging robotics applications. This means
that regulators should understand the attributes that are sought to be
regulated i.e., the entire spectrum of stakeholders and subject matter to
guide them in developing the laws (ibid, 1269). This could potentially
work with other applicable areas of law where care robots are concerned,
as they intersect with other aspects of the law such as liability issues, tort,
insurance, compensation and medical negligence, hospitals liability and
malpractice (Holder et al. 2016, 390). Protecting user safety need not
necessarily conflict with scientific innovations. The European Commission
Robolaw project initiated in 2012 is primarily innovation-driven but
takes a serious view of manufacturers’ responsibilities under existing laws
covering product liability to ensure that consumers are protected from
harm arising from using these products (Holder et al. 2016, 385).
Regulators are expected to regard fundamental principles that apply, such
as the need for robotics not to compromise human dignity, health, safety
and privacy, responsible application of technologies and accountability
for liability where robots caused harm (ibid, 386).
The discussion thus far reveals some feasible options for regulating care
robots. Ad-hoc type regulations may not sufficiently address the purpose
for which care robots are created, (Palmerini et al. 2016) which then
affects the adequacy of these laws in addressing the risks from using care
robots and their liability to harm caused. In arguing for a product-liability
type of regulatory approach for care robots, Palmerini et al. (2016, 80)
drew a parallel to the functions of products similar to existing product
liability laws, thereby precluding special regulatory regimes. It is correctly
observed that liability arising from robotics applications should be regu-
lated in a balanced way that caters to competing interests that are based
on real market considerations rather than assumptions (ibid, 83). This
approach however does not preclude the creation of new customised
rules for certain types of robotics applications, while for other types of
applications, existing laws could be applied with appropriate modifica-
tions. For example, existing consumer protection legislations may still be
relevant where care robots carry risks arising from their design and sale,
however further considerations are needed where they test the limits of
existing consumer protection laws (Holder et al. 2016, 399). It remains
to be seen how this aspect of the law will be applied.
A care robot impact assessment appears to be a feasible option to
address emerging or hidden risks presented by care robots, supported by
active engagements with all relevant stakeholders to feedback and review
the efficacy of laws. The combination of these approaches may mediate
some of the legal difficulties in classifying care robots, as it will become

clear that the purpose and functionalities of care robots are prioritised in
the risk evaluation rather than simply labelling what type of robots they
are. Ultimately in ensuring the safe use of care robots, designers, develop-
ers, and regulators should have joint proactive roles throughout the life-
cycle of care robots, from product design, manufacturing, testing, use
and risk monitoring and management. It is through a holistic, multidis-
ciplinary approach to devising regulatory responses in managing the
implications from using care robots that enables an adaptive and respon-
sive governance framework to emerging risks and concerns.
3 Conclusion
Care robots are continuing to develop according to the needs, contexts,
and use, as well as the level of permissiveness in regulatory landscape
where these applications are developed and deployed. As such it is not
unusual that their developments are fragmented and highly tailored to
the functions they are created to serve. The ethical considerations will
continue to remain relevant, as they raised important implications to the
stakeholders in its lifecycle, from design, production, testing, implemen-
tation and user or market feedback. Research has highlighted the impor-
tance of involving end-users in the design and development processes, a
valid claim as they are the ones who will be interacting with these care
robots on a more frequent basis, and as a result, the effect upon these
users could not be underestimated. The elderly population has vulnera-
bilities which must be addressed in the context in which they use these
care robots. Issues such as control of its use, integration in their daily
lives, their privacy and security are paramount to ensure their safety in
interacting with these care robots. Care robots should meet the purpose
for which they are created, thus regulatory frameworks should be designed
to offer a measure of assurance to these users that their safety and quality
of life are not compromised in favour of scientific advancements.
Balancing these two interests is neither straightforward nor simple, as
they involve careful deliberation of the technology, the end- users, financ-
ing, intellectual property strategies, market needs and developers to guide
appropriate standards while allowing flexibility in innovations in a fast-

evolving area.
Questions regarding commercialisation and intellectual property pro-
tection are increasingly important and any regulatory framework must
respond flexibly to these concerns, particularly the ways in which laws are
supportive of innovations that benefit the population, while seeking to
protect the safety of users. Besides considering the types of regulatory
framework, it is equally important to consider the implementation capa-
bilities of countries in regulating care robots, particularly the prepared-
ness of stakeholders in engaging with the implementation. Researchers
and markets develop in different paces from the experience of end-users,
who are usually perceived as the ‘passive’ recipient of technological appli-
cations and so this risks mismatch of utilities of care robots and why these
are introduced into their lives where human carers are available or if they
preferred human carers. Bringing elderly end- users to the conversation is
a significant first step in ensuring these care robots is designed to function
in a way that serves the interests of elderly end-users.
References
Begishev, I., Khisamova Z., & Vasyukov, V. Technological, (2021). Ethical,
Environmental and Legal Aspects of Robotics. E3S Web of Conferences,
244, 12028 https://doi.org/10.1051/e3sconf/202124412028EMMFT-2020
Caleb-Solly, P., Harper, C., & Dogramadzi, S. (2021). Standards and Regulations
for Physically Assistive Robots. Proceedings of the 2021 IEEE International
Conference on Intelligence and Safety for Robotics. Nagoya, Japan,
March 4–6, 2021
Chalmers, D. J. (1996). The Conscious Mind: In Search of a Fundamental
Theory. Oxford University Press.
Chan, K. Y. G. (2021). Trust in and Ethical Design of Carebots: The Case for
Ethics of Care. International Journal of Social Robotics, 13, 629–645,
https://doi.org/10.1007/s12369-020-00653-w
Chou, Y-h., Wang, S-y. B., & Lin, Y-t. (2019). Long-term care and technologi-
cal innovation: the application and policy development of care robots in
Taiwan. Journal of Asian Public Policy, 12(1), 104–123, https://doi.org/10.
1080/17516234.2018.1492315
Coeckelbergh, M. (2010). Artificial Companions: Empathy and Vulnerability

Mirroring in Human-Robot Relations. Studies in Ethics, Law, and
Technology, 4(3), Article 2. https://doi.org/10.2202/1941-6008.1126
Darwall, S. L. (1977). Two Kinds of Respect. Ethics, 88(1), 36–49. Retrieved
from http://www.jstor.org/stable/2379993
Dickinson, H., Smith, C., Carey, N., & Carey, G. (2022). “We’re Still Struggling
a Bit to Actually Figure Out What That Means for Government”: An
Exploration of the Policy Capacity Required to Oversee Robot Technologies
in Australia and New Zealand Care Services. Int. J. Environ. Res. Public
Health, 19, 4696. https://doi.org/10.3390/ijerph19084696
Festini, S. B. (2022). Busyness, mental engagement, and stress: Relationships to
neurocognitive aging and behavior. Frontiers in Aging Neuroscience, 14,
980599. https://doi.org/10.3389/fnagi.2022.980599
Floridi, L. (2008). Artificial Intelligence’s New Frontier: Artificial Companions
and the Fourth Revolution. Metaphilosophy, 39(4–5), 651–655.
Fosch-Villaronga, E. (2016). ISO 13482:2014 and Its Confusing Categories.
Building a Bridge Between Law and Robotics. In Wenger et al. (eds.), New
Trends in Medical and Service Robots, Mechanisms and Machine Science
(1st ed, pp. 31–44). Springer International Publishing Switzerland. https://
doi.org/10.1007/978-3-319-30674-2_3
Fosch-Villaronga, E. (2020). Robots, Healthcare, and the Law Regulating
Automation in Personal Care. Routledge.
Fosch-Villaronga, E., & Heldeweg, M. (2018). “Regulation, I presume?” said
the robot – Towards an iterative regulatory process for robot governance.
Computer Law and Security Review, 34, 1258–1277.
Fosch-Villaronga, E., & Mahler, T. (2021). Cybersecurity, safety and robots:
Strengthening the link between cybersecurity and safety in the context of care
robots. Computer Law & Security Review, 41, 105528.
Fosch-Villaronga, E., & Özcan, B. (2020). The Progressive Intertwinement
Between Design, Human Needs and the Regulation of Care Technology: The
Case of Lower-Limb Exoskeletons. International Journal of Social Robotics,
12, 959–972 https://doi.org/10.1007/s12369-019-00537-8
Frennert, S., Aminoff, H., & Östlund, B. (2021). Technological Frames and
Care Robots in Eldercare. International Journal of Social Robotics, 13,
311–325, https://doi.org/10.1007/s12369-020-00641-0
Fulbright, R. (2021). A Synthetic Elderly Companion Named Lois. In Q. Gao
& J. Zhou (Eds.): HCII, LNCS 12787, (pp. 403–417). Springer Nature
Switzerland AG.
Garner, T. A., Powell, W. A., & Carr, V. (2016). Virtual carers for the elderly: A
case study review of ethical responsibilities. Digital Health, 2, 1–14, https://
doi.org/10.1177/2055207616681173
Heylen, D., op den Akker, R., ter Maat, M., Petta, P., Rank, S., Reidsma, D., &
Zwiers, J. (2011) On The Nature Of Engineering Social Artificial
Companions. Applied Artificial Intelligence, 25(6), 549–574, https://doi.
org/10.1080/08839514.2011.587156
Holton, R. (1994). Deciding to trust, coming to believe. Australasian Journal of
Philosophy, 72(1), 63–76. https://doi.org/10.1080/00048409412345881.
Holder, C., Khurana, V., Harrison, F., & Jacobs, L. (2016). Robotics and law:
Key legal and regulatory implications of the robotics age (Part I of II).
Computer law & Security Review, 32.
Hoppe, J. A., Melkas, H., Pekkarinen, S., Tuisku, O., Hennala, L., Johansson-
Pajala, R-M., Gustafsson, C., & Thommes, K. (2022). Perception of Society’s
Trust in Care Robots by Public Opinion Leaders. International Journal of
Human-Computer Interaction, 1–17, https://doi.org/10.1080/10447318.
2022.2081283
IEEE. (2019). Ethically Aligned Design: A Vision for Prioritizing Human Well-
being with Autonomous and Intelligent Systems. Retrieved Nov 7, 2022,
from https://standards.ieee.org/wp-content/uploads/import/documents/
other/ead1e.pdf
Iizuka, M., & Ikeda, Y. (2021). Regulation and innovation under the 4th indus-
trial revolution: The case of a healthcare robot, HAL by Cyberdyne.
Technovation, 108, 102335.
Johansson-Pajala, R-M., & Gustafsson, C. (2022). Significant challenges when
introducing care robots in Swedish elder care. Disability and Rehabilitation:
Assistive Technology, 17(2), 166–176, https://doi.org/10.1080/17483107.
2020.1773549
Jones, J. (2004, February 20). Public Gives Bush Benefit of Doubt on National
Guard Service. Princeton, New Jersey. Retrieved from https://news.gallup.
com/poll/10690/public-g ives-b ush-b enefit-d oubt-n ational-g uard-
service.aspx
Juravle, G., Boudouraki, A., Terziyska, M., & Rezlescu, C. (2020). Trust in
Artificial Intelligence for Medical Diagnosis. Progress in Brain Research, 253,
263–282. https://doi.org/10.1016/bs.pbr.2020.06.006
Lancaster, K. (2019). The robotic touch: Why there is no good reason to prefer
human nurses to carebots. Philosophy in the Contemporary World,
25(2), 88–109.
Leiten, I., Pereira, A., Mascarenhas, S., Martinho, C., Prada, R., & Paiva, A. (2013).
The influence of empathy in human–robot relations. Int. J. Human-
Computer Studies, 71, 250–260.
Lim, M.Y. (2012). Memory models for intelligent social companions. In
Zacarias, M., de Oliveira, J.V. (eds.) Human-Computer Interaction. SCI,
396, (pp. 241–262). Springer, Heidelberg.
McTear, M. et al., (2016). Conversational Interfaces: Devices, Wearables,
Virtual Agents, and Robots, (pp. 283–308). Springer International Publishing
Switzerland https://doi.org/10.1007/978-3-319-32967-3_13
Nambu, T. (2016). Legal regulations and public policies for next-generation
robots in Japan. AI & Soc, 31, 483–500.
National Academy of Sciences, Engineering and Medicine. (2020). Social
Isolation and Loneliness in Older Adults: Opportunities for the Healthcare
System. Washington, DC: The National Academies Press. https://doi.
org/10.17226/25663
National Health Innovation Centre. (2022). Singapore Health Assistive and
Robotics Programme Grant. Retrieved Nov 2, 2022, from https://nhic.sg/
web/index.php/collaborations/sharp-projects
Niemelä, M., & Melkas, H. (2019). Robots as Social and Physical Assistants in
Elderly Care. In M. Toivonen, & E. Saari (eds.), Human-Centered
Digitalization and Services, Translational Systems Sciences (1st ed.,
pp. 177–197). Springer Nature Singapore Pte Ltd. https://doi.
org/10.1007/978-981-13-7725-9_10
Palmerini, E., Bertolini, A., Battaglia, F., Koops, B.-J., Carnevale, A., & Salvini,
P. (2016). RoboLaw: Towards a European framework for robotics regulation.
Robotics and Autonomous Systems, 86, 78–85, https://doi.org/10.1016/j.
robot.2016.08.0260921-8890
Pierce, R. (2020). Robots, the Alzheimer patient, and the GDPR: Policy and
privacy considerations for the use of robots in care, treatment, and diagnosis.
Alzheimer’s Dement, 16(Suppl. 10), e043261.
Rantanen, T., Leppälahti, T., Porokuokka, J., & Heikkinen, S. (2020). Impacts
of a Care Robotics Project on Finnish Home Care Workers’ Attitudes towards
Robots. Int. J. Environ. Res. Public Health, 17, 7176; https://doi.
org/10.3390/ijerph17197176
Responsible Robotics. (2022). Assessment principles. Retrieved Nov 2, 2022,
from https://responsiblerobotics.org/quality-mark/assessment-principles/
Sharkey, A., & Sharkey, N. (2012). Granny and the robots: ethical issues in
robot care for the elderly. Ethics Inf Technol, 14, 27–40.
Siau, K., & Wang, W. (2018). Building Trust in Artificial Intelligence and
Robotics. Cutter Business Technology Journal, 31(2), 47–53.
Sparrow, R. (2016). Robots in aged care: a dystopian future? AI & Soc,
31, 445–454.
Sparrow, R., & Sparrow, L. (2006). In the hands of machines? The future of
aged care. Mind Mach, 16, 141–161, https://doi.org/10.1007/
s11023-006-9030-6
Sunstein, C. R. (2015). The Ethics of Nudging. Yale Journal on Regulation,
32, 413–450.
Tan, S. Y., & Taeihagh, A. (2020). Governing the adoption of robotics and
autonomous systems in long-term care in Singapore. Policy And Society,
1–20 https://doi.org/10.1080/14494035.2020.1782627
Van Zaalen, Y., McDonnell, M., Mikołajczyk, B., Buttigieg, S., del Carmen
Requena, M., & Holtkamp, F. (2018). Technology implementation in deliv-
ery of healthcare to older people: how can the least voiced in society be heard?
Journal of Enabling Technologies, 12(2), 76–90.
Van Aerschot, L., & Parviainen, J. (2020). Robots responding to care needs? A
multitasking care robot pursued for 25 years, available products offer simple
entertainment and instrumental assistance. Ethics and Information
Technology, 22, 247–256 https://doi.org/10.1007/s10676-020-09536-0
Vandemeulebrouckea, T., de Casterléb, B. D., & Gastmans, C. (2018). The use
of care robots in aged care: A systematic review of argument-based ethics lit-
erature. Archives of Gerontology and Geriatrics, 74, 15–25.
Vargas, P. A., Fernaeus, Y., Lim, M. Y., Enz, S., Ho, W. C., Jacobsson, M., &
Ayllet, R. (2011). Advocating an ethical memory model for artificial com-
panions from a human-centred perspective. AI & Soc, 26, 329–337.
Von der Pütten, A. M., Krämer, N. C., & Eimler, S. C. (2011). Living with a
Robot Companion – Empirical Study on the Interaction with an Artificial
Health Advisor. ICMI’11, November 14–18, ACM, 327–334,
978-1-4503-0641-6/11/11
WHO. (2022, October 1). Ageing and health. Retrieved Nov 7, 2022, from
https://www.who.int/news-room/fact-sheets/detail/ageing-and-health
Wilks, Y., Bengio, S., & Bourlard, H. (Eds.) (2005). Artificial Companions
MLMI 2004, LNCS 3361, pp. 36–45, Springer-Verlag Berlin Heidelberg.
Wilks, Y., Jasiewicz, J. M., Catizone, R., Galescu, L., Martinez, K. M., & Rugs,
D. (2015). CALONIS: An Artificial Companion Within a Smart Home for
the Care of Cognitively Impaired Patients. In C. Bodine et al. (Eds.): ICOST
2014, LNCS 8456, (pp. 255–260). Springer International Publishing
Switzerland. https://doi.org/10.1007/978-3-319-14424-5_30
An Examination of the Tangible Value
of IP Financing for Companies
and Businesses
Nadia Naim
Intellectual or intangible assets are today recognised by many companies

as their most important resource. Without intellectual property rights
[IPR] many innovative ventures have nothing to sell or licence. In con-
temporary knowledge-intensive economies, from the world’s largest and
most powerful companies to the smallest small to medium sized enter-
prises, the exploitation of intellectual assets; copyright, patents, trade-
marks, designs and know-how are essential to business and the creative
industries. However intellectual property is rarely sufficient, of itself, to
create businesses or, indeed, to create significant economic value, it needs
to be incorporated into a commercialisation plan. A successful future for
IP financing is a significant step in further development of the IP-based
economy. An IP audit is a systematic review of all IP assets (registered and
unregistered) created, owned, used and/or licensed by a business. IP
audits are important to identify how good the processes to capture and
manage IP are. Questions around controlling leakages to customers and
N. Naim (*)
Law and Social Sciences, Aston University, Birmingham, UK
e-mail: n.naim@aston.ac.uk
https://doi.org/10.1007/978-3-031-42576-9_7
158 N. Naim
industry, filling in the gaps in the IP portfolio to support future products

and assets and most importantly, are we protecting the right IP assets in
the right way are all questions to be explored in this chapter.
1 IP and Business

IP occurs through a business and every business owns IP in varying forms
and to various values. Whether starting-up, spinning-out or developing
an existing business, research and development will create volumes of
potentially valuable IP. This could be looking at a brand-new product
from the outset to incremental improvements along the way. Identifying
what has been done at each stage and assessing what IP might exist will
allow a business to decide on whether or not to tie down the IP. Existing
WIPO Treaties, including the Berne Convention, provide the framework
for creators to monetise their creations, however the necessary regulatory
frameworks and mechanism for IP financing unlocking quantifiable
value to grow businesses, have been nascent (WIPO 2021).
Regardless of being a high or low tech business, a brand could become
the most valuable asset a business owns as it is often highly influential to
customers. Whether this is the name of the business itself, or a range of
products being marketed and sold. marketing a business many aspects of
IP will be created, whether this is through TV adverts, social media clips,
blogs or banners or leaflets at exhibitions. Understanding who creates
and owns the rights to what can be important to ensuring the valuable
assets within the business are maintained, to protecting against accidental
infringement on others’ rights, the management of IP rights is a vital
lever to business growth and success (Tang 2022).
IP protects competitive advantage as it creates barriers to entry and
provides freedom to operate, it can support customer lock-in, and as a
result, aligns security with purpose of funds and drivers of cash flows. It
enables scale up for IP rich businesses as unlike tangible assets, intangi-
bles like IP do not wear out with use. In fact, the more widely IP is
adopted, the more valuable it becomes and therefore, when better under-
stood, could be the main focus for company investment. There is already
considerable scale-up expenditure invested in IP rights through research
An Examination of the Tangible Value of IP Financing… 159
and development, software, data, process improvement and training.

Management and investors are well motivated to maintain and retain it,
even if difficulties arise with competitors or in more severe cases, litiga-
tion proceedings (Brassell 2022). Thus, the IP audit and valuation mech-
anisms can enhance a business portfolio by creating an IP business
strategy for maintaining and exploiting its IP rights.
1.1 What Is an IP Audit?
Intellectual property fuels the creative economy, knowledge-based assets

underpin the revenue generated by IP rich companies and could be a vital
competitive edge for start-ups and businesses (Durant 2021). The IP
audit is a systematic review of all the IP assets a company owns, both
registered and unregistered as well as any licencing agreements, to quan-
tify the value that the IP rights add to the overall business finances and
strategy. An IP audit that is too broad or doesn’t align to the business
vision will not be of value. Many companies are able to utilise their IP
value as part of their exit strategy, for example, Cadbury’s sold their busi-
ness for $11.5 billion in 2010. Forty-one percent of Cadbury’s enterprise
value was attributed to its trademark portfolio when the US company
Kraft successfully took over the historic British chocolatiers company.
The extent to which Cadbury secured all of it various IP assets, especially
its trademarks, highlights the value IP can add to the balance sheet and
more importantly, how it is monetised to add significant value to the
company (Millien 2014).
Away from the multimillion pound companies, with the necessary rev-
enue to not only invest in creating their IP rights, but also protecting
them, the focus of this chapter will be on how to better understand the
purpose of an audit for the benefit of the whole business sector. For most
start ups and new businesses, the prospect of protecting your IP can be
exceptionally daunting. What exactly is involved, when is an IP audit
needed, how is it performed and how does an audit inform the overall
business policy and strategy (Singh 2021). Let’s start with the why, IP
gives businesses a much needed competitive advantage, not only against
160 N. Naim
economic free riders, but also as a litigation tool to recover losses and
other civil law remedies.
1.2 Processes to Identify, Capture and Manage IP
An audit allows a company to know what its core assets are, control
mechanisms, actions to strengthen and monetise the company’s IP and
review non-disclosure agreements, confidentiality contracts and IP licens-
ing arrangements. External investment and financing can be challenging
for businesses wishing to expand or avoid an insolvency issue given the
hesitancy in the financial sector to approve financial loans against intan-
gible assets. Land & property are tangible commodities that can be more
easily quantified by lenders and valued against established banking lend-
ing practices. Since the financial crash of 2007, banks and financial insti-
tutions are under stringent monitoring and are generally averse to
non-conventional risk in asset-based financial lending. However, where a
company can demonstrate it has protected its intangible assets and the
competitive advantage gained through patents, copyright, and confiden-
tiality agreements, are adequately protected and accounted for in an IP
audit and it can be demonstrated that they underpin revenues and fore-
casts, then banks or investors will consider taking security over them. The
diligence-backed IP audit can demonstrate how a company’s IP portfolio
clearly underpins its products & services and drives the financial growth
of the business. Further, where the IP audit incorporates risk assessments
for the level of protection for current and upcoming products and ser-
vices, the higher the likelihood of success with investors and financial
institutions. When possible recovery value can be identified from licens-
ing and sale of assets, and subsequently implemented, lenders are more
reassured that risks have been mitigated as part of the exit strategy.
Investment in IP development for growing a business can be costly
especially where there are different divisions or business streams. As
attractive as successful innovation and entrepreneurship is to all busi-
nesses, there is a significant investment needed in the form of research
and development which more often than not, can have a costly number
of failures before successful IP rights are created. Businesses rely on
Intellectual Property valuations to make informed decisions about

Intellectual Property protection, return on investment on marketing,
research and development, mergers, acquisitions and licensing to estab-
lish to lenders that the company has protected the relevant intellectual
property assets that are core assets for the securement of future cash flows
(UK IPO 2014).
1.3 Contents of an IP Audit
An audit allows for a review of all confidentiality policies to make sure

that they are relevant and appropriate. Identify what IP relates to each
product and whether the assigned IP right will last the lifetime of the
product and capture new products that are in the development stage.
Along with this, the contents must cover the likelihood of infringement
and robust monitoring mechanisms for the llikelihood of discovering
infringement. The audit will establish whether the appropriate IP policies
and procedures are in place for the IP rights identified, their validity and
status. Identifying the most valuable rights with who is responsible for
the corresponding IP policies are needed to further the IP strategy in line
with the business plan and to decide what assets should be protected.
Procedures are needed to ensure that all IP rights are logged, maintained,
and protected. This could be through licensing agreements and legally
binding terms that are to be adhered to.
Starting with registered IP assets, primarily, patents, designs and trade-
marks, a checklist is created to identify a full list of all registered and
pending assets, countries in which they are or will be registered, a record
of all maintenance fees and dates of renewal. For patents and design
rights, a record of inventors and designers with filing dates and numbers,
status and expiry date. The IP audit checklist should identify product
lines and brands to which each is relevant. (Bainbridge and Howell
2014). A fit for purpose IP audit covers, the management of IP assets,
commercialisation, ownership of IP rights, how best to harvest IP, licens-
ing and franchising agreements, insurance, valuation, tax relief advice
and fees for IP services in the UK and abroad.
162 N. Naim
1.4 Patents Audit
The first step is to identify what patents the business has by listing all
registered and pending assets. This includes countries in which they are
registered or will be registered. Patent registration is the most expensive
of all the IP assets and as such requires a patent attorney. For the IP audit,
patents that are filed and/or approved need a record of all fees and details
of the inventors or designers. Filing dates and numbers with status and
expiry date can be aligned to identifiable product lines to which each
invention or design is relevant to. Patents are international in nature, and
are covered by TRIPS (Articles 27 and 28). In the UK, A full patent
application under Patent Act 1977 must contain a request for the grant
of a patent and an abstract which gives technical information, in the form
of a specification about the invention and the field to which it contrib-
utes. The function of the invention needs to be clear and precise to grant
a monopoly right (Strix Ltd v Otter Controls Ltd [1995]). A priority date
is then given as the date on which an application is filed and the duration
of the patent is calculated from that date. The UK patent will only apply
in the UK and an analysis of the territorial reach of the patent is needed
as part of the audit determine whether international protection is neces-
sary and of value.
The patent process involves a series of steps; starting with the applica-
tion form and the request for a search from the UK IPO, the Preliminary
examination is regulated by section 15A, PA 1977 with an 18 month
turnaround window, publication is 18 months from the priority date,
followed by Substantive examination, a period for amendments and if
successful, the IPO grants the patent. The purpose of the audit is to con-
sult relevant patent databases and create a list of any competing patents
that are used by third parties.
1.5 Design Rights Audit
Design rights in the UK can be registered or unregistered. Design rights

that can be registered, share some features with patent law, in a very broad
sense, and applies to designs that have or are intended to have eye-appeal.
These designs can be described as aesthetic. The other system of design

protection is called the design right and is provided for along copyright
lines. This right applies to designs that can be said to be functional in
nature such as a new design for an engine cover or a plastic printer ink
cartridge. It can, however, also apply to many registrable designs and
there is a large overlap between the two forms of design rights. Both
forms of design right relate to the design aspects of the shape or configu-
ration of an article and, for registered designs only, also to pattern and
ornament.
A design is the outward appearance of the whole or parts of a product.
A product can be any industrial or handicraft item. Examples of design
features include lines, colours and shapes. Designs must be novel or new
and no other identical design has been made available to the public. The
design needs to have individual character. Important for the audit check-
list is to establish whether the same ‘overall impression’ on the ‘informed
user’ has already been disclosed. Some designs are excluded from protec-
tion by law because they run counter to public order and morality, and
exclusions need to be checked. When a product must be a certain shape
to perform a function, then it does not have design freedom, i.e. the part
of a key that fits into the lock that it was designed with.
The unregistered form of the design right works similarly to copyright
and is discussed below. Keeping a record of owning the right, who has
designed it and when, is evidential proof for owning the right, whereas
with a registered right you have a certificate to provide evidence that the
right exists. Design rights have significantly shorter terms and criteria for
protection, dependent upon time from marketing and creation (UK IPO).
1.6 Trademarks Audit
Article 3 of EU Trade Marks Directive removed the requirement of

graphical representation (Regulation (EU) 2015/2424). Integral to the
trademarks audit is to ensure a registered trade mark or upcoming trade
mark consist of a sign, in particular words, including personal names, or
designs, letters, numerals, colours, the shape of goods or of the packaging
of goods, or sounds, provided that such signs are capable of: (a)
164 N. Naim
distinguishing the goods or services of one undertaking from those of

other undertakings; and (b)being represented on the register in a manner
which enables the competent authorities and the public to determine the
clear and precise subject matter of the protection afforded to its propri-
etor. The representation needs to be “clear, precise, self-contained, easily
accessible, intelligible, durable and objective” (Section 13 in the pream-
ble of the Directive).
The application process can be handled internally or with a trademark
attorney. A request for registration can be made to the relevant IP office,
with information about the identity and address of the applicant, state-
ment of goods/services which are the subject of the trademark registra-
tion, representation of the mark and the statement will be used. The
applicant receives a filing date, the registrar examines the application
against absolute and relative grounds for refusal (Trade Mark Act
1994, UK).
Section 3 of the Trade Mark 1994 Act gives the absolute grounds for
refusal (derived directly from Article 3 of EU Directive 89/104/EEC),
and sets out categories of signs which cannot be registered. However,
signs in 3 of these categories are capable of registration if accompanied by
proof of secondary (or acquired) meaning.
The 3 categories of marks which cannot be registered unless they show
proof of a secondary or acquired meaning to give a distinctive character
as a result of use are:-
S.3(1)b – Trademarks which are devoid of any distinctive character.

S.3(1)c – Trademarks which consist exclusively of signs or indications
which are descriptive of the goods or services themselves.
S.3(1)d – Trademarks which consist exclusively of signs or indications
which have become generic terms in general language or the specific
market sector. (Trade Mark Act 1994. UK).
When a trade mark application is examined at the Intellectual Property

Office, it may attract objections based on one or more of these categories.
The types of marks that can fall into the above categories are words/signs/
devices/shapes which are wholly descriptive of the product.
The registrar will also examine existing marks, and trade marks may
also be refused registration on the basis of relative grounds. This is con-
cerned with the effect of registration that the proposed mark may have on
other pre-existing marks. It used to be the case that the UK Trade Mark
registry raised these grounds at the application stage. However, it is now
left to owners of trade mark registrations to oppose applications through
publication in the trademark journal. The trade mark application is then
open to public inspection and possible objection for a period of three
months from the date of publication.
If, as occasionally happens, an objection is raised by a third party
against the application, then the matter has to be considered and appro-
priate action taken. It may be possible to deal with such an objection in
a relatively simple exchange of correspondence. On other occasions, the
ensuing opposition proceedings can be long and complicated.
If there is no objection, or any objection has been resolved, then the
application can proceed to registration on completion of certain formali-
ties at the Trade Mark Registry. A registration certificate is then issued
and the trade mark is placed on the Register for a period of ten years from
the date of application. After that, it can be renewed indefinitely.
For classifications of the trade mark, the Nice classification system can
be utilised to ascertain the classes of goods for which this mark should be
applied and nationally, IP offices have a trade mark search to allow for
comparison with existing classified marks. (Trade Mark Search, UK IPO).
2 Passing Off
As well as trade mark protection, the audit benefits from an understand-
ing of brand power and how to protect unregistered marks. Brand power
accumulates significant goodwill which can enable businesses to success-
fully extend the brand to new products, gain trade leverage in the com-
petition for retailers’ distribution points and take advantage of potential
cost savings in the promotion of new products in comparison to potential
new entrants or competitors with less brand power. In addition to the
brand’s reputation amongst consumers,
166 N. Naim
Despite the availability of statutory protection for brands under the

Trade Marks Act, the common law tort of passing off remains important,
for at least two reasons. First, many small businesses are not equipped or
aware that they should register their trade marks early, and as a result,
may be forced to rely on passing off. Secondly, a mark which is otherwise
invalid or unregistrable, as discussed above on absolute and relative
grounds for refusal, may be able to gain some level of protection through
passing off.
Passing off has recently proved useful in combatting ‘cybersquatting’,
the practice of registering Internet domain names containing company
names or trade mark names by persons hoping to sell the names at grossly
inflated prices to the relevant companies or trade mark proprietors. As a
common law tort, it aligns to the concept of unfair competition. It is
recognised in international law under the Paris Convention for the pro-
tection of Industrial Property.
3 Measuring the Value of Goodwill

In IRC v Muller & Co’s Margarine Ltd [1901], goodwill was defined as
the attractive force which brings in custom, or “whatever adds value to a
business”. Goodwill can be seen as value added elements which lead con-
sumers to purchase a trader’s goods/services. It is not the same as reputa-
tion, as reputation can occur without a consumer purchase. Consequently,
goodwill can be seen to exist in company/product names, devices/logos,
and the particular way in which a product is packaged. Goodwill needs
to be established by evidence. One of the main issues in establishing
goodwill is its locality and scope. For example, a restaurant with an excel-
lent reputation for food is fixed in one location, any goodwill it achieves
is unlikely to be nationwide. However, companies have been able to show
goodwill even if they do not trade in a particular locality, region or even
country (Sheraton Corp. of America v Sheraton Motels [1964]).
To quantify the goodwill, a business can satisfy goodwill by demon-
strating a reputation and a customer base (Countess Housewares v Addis)
and it can also exist in things such as a business name, devices and its
get-up. The case of Coca-Cola v Barr found in favour of Coca-Cola and
held the style of Coca-Cola’s bottle was part of their get-up and subse-
quently demonstrated their goodwill. Together with goodwill, the tests
for passing off require a misrepresentation and consequential damage
to occur.
3.1 Hidden Assets
Hidden assets can include know-how, trade secrets, databases, mailing

lists, customer lists, contractual agreements, creation of websites, domains
and training programmes. For trade secrets and confidential information,
one named person should be in control of all relevant information, and it
should be recorded who this person is. They should be tasked with iden-
tifying, grading, tagging, and giving sensitive information an appropriate
level of protection. Confidential information should be classified accord-
ing to its importance to the business and each grade should have polices
dealing with access and sharing of that type of information, and it must
be stored appropriately to retain its status. Employees must know what is
regarded as a trade secret and confidentiality agreements can be included
in contracts of employment. Accurate and detailed contractual clauses
can stipulate the importance of adhering to the confidentiality policies
and understanding the consequences of breaching the terms of the con-
tract. The law of breach of confidence developed in equity as a way of
protecting confidential information by preventing its use by employees to
whom the information has been divulged in confidence or avoid further
unapproved disclosure. Highly confidential information should be
restricted on a need-to-know basis, it would need to be stored safely and
password protected, with secure records kept.
3.2 Copyright Audit
Given that copyright is an unregistered IP asset, for auditing purposes, a

copyright policy will be the central document to accurately record and
protect all copyright works. Separate works can be categorised dependent
upon the core copyright works for the business and could include
168 N. Naim
software, web pages, training materials and advertisements. Meticulous

record keeping is required for unregistered assets and the copyright policy
can include time sensitive information. For example, the date the work
was created and where, determines the length of protection and provides
evidence if there is a dispute or suspected infringement. Linking back to
the employment contract, any copyright created by employees would
need to have a valid clause in place as to who owns the work. This would
determine who is named as the author with information as to their status.
If commissioned, the authorship would most commonly be retained by
the company and any assignments or licences of the copyright to be
agreed with the necessary terms (Copyright, Design and Patents
Act 1988).
4 IP Due Diligence

Once the IP audit is complete and there is a clear record of all IP assets,
IP due diligence can provide a snapshot of where the IP assets and liabili-
ties, in relation to an IP transaction, are identified and verified. It is nar-
rower in scope than an audit as it relates only to the IP rights that are
relevant to the transaction. It is tailored to an event, such as purchasing
of a company or investment purposes. Further, it can involve buying or
selling IP, litigation, filing patents, the development or launch of a new
product or as an exit strategy.
Key considerations are drawn from the IP audit and an assessment of
the authority to use, the validity, duration, strengths and weaknesses of
the rights, the vulnerability to challenge and could a competitor use their
own IP to block your freedom to operate, potentially preventing the
product being exploited and as a result, diminishing the value of the IP
rights. The question of whether there is any prior art poses many issues,
it is impossible to prove that something does not exist, but a thorough
search gives some confidence to mitigate the risk. If there are gaps in the
IP due diligence process, potential purchasers or licensees will require
warranties or indemnities to protect their investment and minimize the
risk as the aim of the due diligence is to minimise the chance of defects
and the transaction falling through (Bainbridge and Howell 2014).
4.1 How Does IP Generate Revenue?
Intellectual property rights grant the owner a monopoly right for a set
period; this allows the owner to make commercial gain from their inno-
vation in return for sharing it. In practice, for example, the existence of a
patented breakthrough product can give a business a head start over their
competition. The improvement of existing products or brand recognition
can give a business the edge and allow them to charge a premium cost on
their IP protected products. Intellectual property, as it states, is a prop-
erty. It is an asset of the owners, to do with as they wish. Unless using IP
to create brand value or using the monopoly to develop a market share a
business may decide to sell or license their assets.
Franchising is a version of licensing where the owner provides support
and the rights to use certain assets to anyone who can pay the fee and
agrees to the terms. Subway®, McDonalds® and Europcar® are all examples
of franchising business models. Businesses can also use IP to cross license,
allowing competitors to use their assets in return for the right to use the
competitors assets, and gain access to other technology that may other-
wise cost money, or even be inaccessible.
Looking at the international route, when and if a business decides to
expand beyond the national level of protection, IP rights can be vital to
ensure the business can have access to the wider market and assert its IP
rights. Having the appropriate rights can help to gain key partners in the
countries that a business is looking to trade in or establish a foothold in
the market (UK IPO).
5 IP Valuation Approach

There are costs involved at all stages of the IP cycle and as such business
are becoming more aware of the need to value their IP. The cost element
of IP begins at the development stage and achieves full life cycle stage by
acquisition. Factoring in the cost of IP generation, allows business to
bring the cost of IP rights onto the balance sheet as it can account for
development costs at a capitalization stage and businesses can spread the
170 N. Naim
investment made in some IP assets over the expected period of benefit,

although there are many restrictions and limitations. For the acquisition
stage, accounting rules are required to quantify the value of the assets
after a business has been bought, as the acquirer has to decide what it
paid for the registered IP assets, and unregistered hidden assets such as
goodwill. Several challenges arise because of valuation rules, for example,
the value of core assets may seem to be going down, when it may in fact
they may be going up. Accurate, regular review of IP value and auditing
avoid having a situation where IP assets appear on the balance sheet at a
point of sale and as an exit strategy only due to the difference in rules
between accounting for development and accounting for acquisitions.
(Brassell 2022).
Valuing IP is not as easy as tangible assets and as such requires a valu-
ation method that is suitable and relevant to the industry the business
operates in. Intellectual property rights change in value for a variety of
reasons and a regular periodic review of IP rights can help quantify the
current value of the rights. For example, the worth of a brand name can
vary significantly after years of marketing, likewise a protected patent can
only hold value and generate revenue if the product stays relevant and
avoids becoming redundant. A patent can start with compelling potential
and begin its life as a unique solution to a problem, but in time other
solutions to the problem may be found which reduce its worth.
Alternatively, successfully marketing your product can ensure your patent
is very valuable. Trade marks generally gain value as they become better
known. The stage of development of the IPR, the availability of informa-
tion and the aim of the valuation all have a bearing on the method used
and each will be discussed next (UK IPO).
There are three main standardised approaches that are used in intel-
lectual property valuation, it is currently not a regulated field, therefore
different valuation experts have different approaches (Calboli and
Montagnani 2021). The following approaches that are used for intellec-
tual property valuations are provided by the UK IPO as income based,
cost method and market value based.
5.1 The Cost Method
The cost method is a value estimate and is cost adjusted for depreciation
and obsolescence. It is based on the aim of establishing the value of an IP
asset by calculating the cost of developing an identical or similar asset
either internally or externally. The cost method poses two fundamental
questions. First, how much would it cost to reproduce a given set of
assets, taking into account the full cycle of asset production from research
and development to acquisition. Second, how much would it cost to
replace a given set of assets and it is within these constraints that the cost
method values IP assets. The method attempts to determine the value of
an IP asset at a particular point of time by aggregating the direct expen-
ditures and opportunity costs involved in its development and mitigating
for redundant IP assets. Costs can include labour, materials and equip-
ment, research and development, overheads for utilities, accommodation
and support staff, creating a prototype, testing and trials, regulatory
approval, certification, and registering the IP asset.
To reiterate, the more detail the IP audit, the more significant the bear-
ing it has on the cost method and what a potential buyer can expect to
pay for avoiding the same costs, instead opting to buy the IP assets.
Valuable benefits that can make the cost method more attractive is the
saving of time for the purchaser, quantifiable expenditure which reflects
the cost of attempting to recreate the same IP, and there is already a track
record of success with the IP assets. Most importantly, the IP assets are
already protected which not only limits other options but also prevents
IP protection for the similar assets given the remedies available against
infringement. This method of valuing intellectual property assets lends
itself to an overall assessment when buying a business. However, as the
emphasis is on costs, rather than profit, it can skew the figures so that
market potential is not fully appreciated. The cost method does not take
account of future value and therefore loses out on a standard by which
value is traditionally calculated.
The approach of calculating replacement costs for IP asset with an
equivalent asset of similar use and function, can be used more reliably
when considering a going concern and solvent valuation. The
172 N. Naim
corresponding IP audit would need to have detailed information relating

to the full value historic costs, which evidences the cost of the IP assets,
including those still in development. Intellectual Property assets that
have future value and are expected to provide an economic benefit, are
better suited to the income-based approach, which is considered next
(Calboli and Montagnani 2021).
5.2 The Income or Economic Benefit Method
The income-based approach can be a very useful method as it values the

revenue the IP assets can generate based on the amount of economic
benefit that the intangible asset is predicted to make, adjusted to its
present-day value. It is a popular method that anticipates the costs and
benefits over the expected lifespan reliable and incorporates externally
verified market information. The net profit value is the sum of the differ-
ence between benefits, less costs, with an allowance for the discount rate.
This method of valuation is based on an assessment of likely future events
and is used to consider investments based on whether the net profit value
is positive or negative. Central to the income based method is what con-
tribution do the IP assets make to current and future sales or profits, what
are the cost savings that can be attributed to specific assets and what is the
discount rate to take account of how the money spent could have been
made otherwise. Calculating the net profit value is complex. Estimating
the economic life of the IP assets over many years requires an estimation
of the benefit year on year, with appropriate weighting depending upon
the industry. The weighting of the benefit of a pharmaceutical drug will
be weighted to the end of the patent whereas fast fashion is much more
near term loaded. Therefore, an assessment of how the value and benefit
will vary over time is needed as value from IP can be hard to isolate,
future projections can be uncertain, especially if market traction is lim-
ited. On the plus side, it is a strong incentive for investors to purchase IP
assets that have potential as if the expected future IP value is positive, it
can be a very attractive investment.
There are several factors that impact on the income based approach,
such as, the strength of the IP assets, market size, the industry, the nature
of the competition, changes in the economic climate and the cost of reg-
istering, enforcing and defending the IP assets. The way in which the IP
assets are exploited, the costs involved, the time it will take to get to mar-
ket and the risks involved all influence this method. A sub method of the
income or economic benefit method is the relief from royalties method
and is based on an assessment of what royalty costs a company is avoiding
by virtue of owning the IP.
Alternatively, the relief from royalties can act as a security for borrow-
ing money from lenders. In 1997, David Bowie issued a 10-year asset-
backed bond on the value of his future royalties from publishing rights
and master recordings from 25 pre-recorded albums and raised
US$55 million. The purchaser received the right to future royalties from
the albums until the principal plus 8% annual interest had been repaid
(WIPO 2016). Securitisation allows a business to bundle IP assets and
lend against predictable future cash flows directly attributable to the
IP. The loan size can be determined by the net present value of the
expected future profits and gives a manageable discounted cash flow
approach, that is adjusted to reflect the risk. There are also opportunities
to utilize IP assets as collateral, this is usually relevant to underwrite a
recovery value, with a discount rate for going concern expectations
(Inngot 2022).
5.3 The Market Value Approach
The market approach is based on a comparison of the actual price paid

for a similar IP asset under comparable circumstances, such as, previous
transactions. Using this approach for IP valuation can be challenging as
assets tend to be unique and novel by nature, and the accuracy of the
valuation involves an analysis of the market, an exchange of IP assets, or
a group of comparable or similar assets, and where the assets are not com-
parable, variables are essential to control for the differences. The key
advantage to the market approach is that it mirrors tangible asset financ-
ing by quantifying the value of the IP assets from prior factual transac-
tions. The business doesn’t have to put its assets on sale to work out what
people will pay. Rather, the approach examines whether comparable
174 N. Naim
assets have been sold in the market, followed by analysing data on how
many businesses, owning comparable assets, have been sold and whether
the investment data reveals how much businesses owning comparable
assets are worth and what is the licensing agreements value on these assets
(Calboli and Montagnani 2021).
Therefore, calculating the value of IP assets by reviewing the sale or
licensing of similar products in the market can act as a benchmark for
estimating the value of a product based on its existing track record in the
market. However, in practice, finding publicly published data on IP
transactions, is very difficult as there is no legal requirement to make such
data publicly available and therefore, finding relevant data that can be a
source for a market approach, is very difficult. Even where there is pub-
licly shared data or a bilateral sharing agreement, the market approach
can still be fairly broad as very few IP transactions will provide a valid
comparison. The IP assets may differ on exclusivity rights, territorial dif-
ferences, market conditions, payment structures and the value from spe-
cialist support.
Of the three, it is the most difficult approach to apply to all of a com-
pany’s IP assets as it isn’t suitable for all types of IP. For example, a patent
is granted for a new or novel invention. Finding comparable information
for a novel invention, by its own description, will be a challenge and an
objective standard is applied to create financial projections that can still
reflect the market value through consumer analysis data. For other IP
assets, such as copyright, the market approach method is very effective as
royalty rates in each market sector, can be benchmarked as high, medium
or low (UK IPO).
5.4 Exploiting Your IP
Exploiting the IP a business owns can create new economic revenue and
achieve other strategic goals through commercialization. IP assets can be
listed on the company accounts and can be a commodity that is bought,
licensed, franchised, or sold. IP assets can be used to secure funding from
banks and lenders, as a fixed or floating charge, depending upon the rel-
evant IP due diligence and audit checks as to who controls the asset. The
main feature to establishing a valid fixed charge over an IP asset is giving

the lender control over the asset. For registered IP assets, with appropriate
provisions in the security documentation, such as documenting the secu-
rity and lender control on the relevant IP register. If there are no control
mechanisms in place, the security will instead be potentially granted as a
floating charge (Popple 2020).
There are several options to IP commercialization from assignment,
licensing, franchising, collaborations, joint ventures, diversification,
sponsorship and negotiated access to technology with third parties.
Assignment is the most straight forward option as by selling the IP asset,
the business is selling all of the attached rights to the IP. Assignment can
generate immediate cash and the business is no longer in control of the
IP right, and as a result would be infringing their own IP creation by
using it, post assignment, without a license.
Under a license agreement, the licensee and licensor are contractually
bound to a set of permissions offered to the licensee for an agreed price to
the licensor. According to Licensing International’s Annual Global
Licensing Survey, global sales revenue generated by licensed merchandise
and services grew to $315.5 billion in 2021, and global industry royalty
revenues reached $17.4 billion. Through a license, an IP owner contrac-
tually enters into a legally binding agreement with the licensee and grants
the licensee rights to exploit some or all aspects of a particular IP asset. In
return, the licensor normally receives compensation as consideration, or
a combination of compensation and equity in a business. The purpose of
licenses is to create the optimal opportunities to exploit the IP assets, and
make the most efficient use of the IP, increasing the economic value and
profit that can collectively be achieved by the license agreements
(Contreras 2022).
Licensing can be attractive for several reasons, not least of all because
of the size of the industry. Often, the licensee has established markets and
suppliers, companies can licence-in and there is no transfer of ownerships
as the IP ownership is still retained by the licensor. It can also be advanta-
geous over franchising agreements as once the transaction is complete,
the level of involvement for the licensor is minimal. There can be upfront
payments as well as royalties without losing all rights to the use of the IP
asset and is commonly associated to research collaborations with direct
176 N. Naim
access to funds and equipment. Licensing agreements can lead to good

links to commercial collaborators and access to markets where the licensee
has experience and means of penetrating the market. The licensor still has
control as once the license has expired, the full control over the IP asset is
back with the licensor.
5.5 IP Financing and Emerging Technologies
As the space inhabited by artificial intelligence continues to grow and

transcend aspects of intellectual property creation, there could be room
for AI-assisted and AI-generated systems to add to the IP audit and IP
valuation methods. As discussed earlier, there are three standardized
methods of IP valuation, each with its own advantages and disadvan-
tages. Emerging technologies can assist with both the auditing and valu-
ation systems and help address the different priorities and approaches. AI
systems can be developed to enhance IP valuation, overcome buyer and
seller disincentives. For the seller, there will be concerns that once core
assets are sold, the remaining assets may not be of interest or if the IP is
all sold to another business, what will be left of the company. Alternatively,
for the buyer, concern can arise over how sustainable the IP assets are or
do they rely on a specialist team and it can be difficult to decide on what
the value is without comparable transactions. In order to evaluate the role
AI systems can play in IP financing, the broader legal landscape pertain-
ing to data to creating common data spaces in the IP industry, from
which to generate more accurate IP comparable data and support IP
financing and lending (Williams 2022). For AI, there is potential to over-
come its challenges around ethics and trustworthiness and create AI
assisted and or generated valuation methods. The UK IPO offers an IP
finance toolkit which acts as a guide for businesses and supports their IP
audit before applying for an IP backed loan or financing. To further sup-
port IP management and commercialisation strategies, as well as raise
awareness of finance options available, toolkits can be useful however AI
systems that can address the limitations with current valuation processes.
6 In Conclusion: IP Financing for Businesses

IP audits can be very insightful and relevant to a company’s financial
forecast and value. The audit will establish whether the appropriate IP
policies and procedures are in place and what IP rights, a company has.
The more detailed and accurate, the more reliable the audit is, in terms of
validity and status. The key priority of the audit varies from the identifi-
cation of the most valuable assets and the associated product lines, as well
as how well they are properly protected. The IP audit is periodically
reviewed to account for licences, terms and conditions, pending actions
and legal proceedings. Cost of development of IP rights is the only way
to bring IP assets onto the balance sheet. The audit will establish whether
the appropriate IP policies and procedures are in place, and record regis-
tered or unregistered assets, their validity and status. The audit aims to
identify the most valuable rights as core assets. IP due diligence is critical
for identifying and verifying IP assets and liabilities, in relation to an IP
transaction.
The IP audit and due diligence of all relevant IP assets, including reg-
istered assets such as patents and trademarks and also unregistered assets
that include copyright, trade secrets and know-how, all contribute to the
valuation of a company. To fully exploit the IP a business owns, the three
different valuation methods monetise IP and generate revenue. For the
cost approach, the valuation method quantifies the reproduction cost of
a given set of assets from inception to production and what the cost of
replacing the IP assets would be. As a methos, this is the most straightfor-
ward reflection of the investment made by the business in terms of tan-
gible data from the cost of human capital, direct costs, overheads, and
other ancillary costs to create a cost-based valuation of the IP assets. The
main disadvantage is that although cost does reflect the investment made
and the skill, labour and capital that has been spent on creating the IP
assets, that doesn’t necessarily mirror the value and therefore is often the
least attractive of the valuation methods.
The market value method is focussed on the comparable value of the
assets. This is an analysis of both comparable assets that have been sold on
the market and the sale of businesses owning comparable assets. The
178 N. Naim
market value method relies more on investment data on the IP value and
ancillary licensing agreements. Despite this method having the most
resemblance to tangible asset valuation, it is flawed as the market com-
parison of intangibles are not as readily available as physical property. The
final method of income-based valuations is the most popular of the three
methods as it assesses the value of the IP assets based on current and
future profits. On the other hand, it is uncertain and can be difficult to
mitigate the risk, if the market traction is limited. To further complicate
matters, valuation priorities are different for each approach and depends
on the purpose of the IP audit such as securitisation, collateral, licensing,
franchising, or equity financing. AI can play a very important role in the
development of AI-assisted and AI-generated audit and valuation meth-
ods to expand the current market limitations on ascertaining the most
accurate value of IP assets and creating more opportunities for IP financ-
ing and lending.
References
Bainbridge, D and Howell, C. Intellectual Property Asset Management : How
to Identify, Protect, Manage and Exploit Intellectual Property Within the
Business Environment, (2014), Routledge.
Brassell, M. Challenges in harnessing IP for finance, (2022), Inngot Limited.
Calboli, I. and Montagnani, M. Handbook of Intellectual Property Research:
Lenses, Methods, and Perspectives (2021), Oxford University Press.
Contreras, J. Introduction to Intellectual Property Licensing. In Intellectual
Property Licensing and Transactions: Theory and Practice, (2022), Cambridge
University Press.
Copyright, Design and Patents Act, 1988
Durant, I. Unlocking potential of intellectual property rights to support the
creative economy, (2021), United Nations Conference on Trade and
Development.
Millien, R. The use of competitive intelligence in IP monetisation, (2014),
Intellectual Asset Management.
Patent Act 1977
Popple, L. Security over IP. Considerations when taking security over IP, (2020),
Taylor Wessing.
Singh, J. How startups and SMEs should think about IP: an investor’s perspec-
tive, (2021), WIPO Magazine.
Tang, D. Inaugural High-level Conversation on Unlocking Intangible Asset
Finance, (2022), World Intellectual Property Office.
Trade Mark Act 1994
UK Intellectual Property Office
UK Intellectual Property Office (UK IPO), IP Finance Toolkit, (2014). Available
at IP Finance Toolkit (publishing.service.gov.uk).
Williams, L. Anonos Secures $50 Million in IP-Backed Financing to Deliver
Data Privacy Technology with 100% Accuracy and Utility to Data-Driven
Enterprises, (2022), Business Wire.
World Intellectual Property Office, (WIPO), Methodology for the Development
of National Intellectual Property Strategies, (2016). Available at https://www.
wipo.int/edocs/pubdocs/en/wipo_pub_958_3.pdf
World Intellectual Property Office, (2021), Unlocking IP-backed Financing:
Country Perspectives. Singapore’s Jouney. Geneva: WIPO.
Transformative (Bio)technologies
in Knowledge Societies: Of Patents
and Intellectual Commons
Mariela de Amstalden and Nivita Sukhadia
1 Transformative Biotechnologies:
Cell-Cultivation for Human
Food Consumption
The year was 1931. In a line perhaps more suitable to science fiction writ-
ing at the time, Winston Churchill envisioned cell-based meats by declar-
ing that: “[w]e shall escape the absurdity of growing a whole chicken in
order to eat the breast or wing, by growing these parts separately under a
suitable medium” (Churchill 1931). Arguably, a technology, and the
products it produces, is only disruptive when it reaches the masses.
Promoted as a more sustainable alternative to conventional proteins,
research into cell-cultivation technology for human food consumption
(cellular agriculture) has gained new urgency in response to pressing
M. de Amstalden (*)
Law School, University of Exeter, Exeter, UK
e-mail: M.De-Amstalden@exeter.ac.uk
N. Sukhadia
University of Cambridge, Faculty of Law Cambridge, Exeter, UK
https://doi.org/10.1007/978-3-031-42576-9_8
182 M. de Amstalden and N. Sukhadia
global challenges. In essence, these products are produced by tissue and

bioprocess engineering that result in a product that is molecularly identi-
cal to conventional agricultural ones, like meat.
In light of this, some initial regulatory activity with the aim of accom-
modating (or hindering) the placing on the market of such products could
be identified in different jurisdictions, most notably in Singapore -the first
country to grant regulatory approval for cell-cultivated meats. Nonetheless,
the regulatory pushback from established global market players in the
meat, fish and diary industries is noticeable. For cell-cultivation food com-
panies, their new products provide the most efficient answer to reduce
greenhouse gas emissions, feed a growing world population and prevent
food-borne diseases. For conventional producers of food, cellular agricul-
ture products are artificial and thus, not equivalent. In our highly inter-
twined global food supply chain, biotechnological advances tend to
precede existing rules-based regulation, challenging the effective respon-
siveness of a valid legal framework. I have argued elsewhere that the lan-
guage used within the law will thus undoubtedly have social, economic,
political and cultural consequences for the future of food systems (de
Amstalden 2021), but also for future global governance mechanisms (de
Amstalden 2016).
Conversely, decades of industrialised animal farming have significantly
contributed to anthropogenic climate change. Livestock-based meat pro-
duction limits the use of scarce natural resources like water and land,
while also releasing large amounts of Greenhouse Gas (GHG) emissions.
The amount of animal protein required to feed a global population that
is expected to reach 9.7 billion in 2050 is likely to double, compounding
the severity of implications for the sustainable use of available resources.
While emerging biotechnologies are currently being developed to mit-
igate the negative effects of livestock-based meat consumption, the lack
of a responsive regulatory framework has the potential to stifle innova-
tion, just as uncertainties about likely technical and social outcomes
inhibit informed decision-making. Arguably, effective stewardship of
cell-cultivation technology products requires broad political and social
assessment and engagement.
Intellectual Property Rights (IPRs) have been inextricably intertwined
with the emergence of transformative technologies as means to reward
Transformative (Bio)technologies in Knowledge Societies… 183
intellectual creation since the beginning of human ingenuity. Cell-

cultivation technology to produce food for human consumption is no
exception. From trade secrets to patents and trademarks, the production
of food for human consumption using cell-cultivation technology creates
a novel scientific field (‘cellular agriculture’) while confronting innova-
tors, manufacturers, regulators and consumers with an array of chal-
lenges. What we understand as IP, and how we engage with it, will shape
the contours of academic discourse, inform public policy debates deter-
mine entrepreneurial success and challenge consumer perceptions.
Encountering IP law at a multiplicity of levels, cellular agriculture as an
emerging field of (legal) enquiry that appears to challenge Lockean
approaches to IP as a legal monopoly, questioning their limits to promote
social progress.
2 Saving the Intellectual Commons

in Future Knowledge Societies
By its very nature, transformative biotechnologies confront the law with
an array of issues, from responses to risks in light of scientific uncertainty,
labelling and consumer protection, restrictions on international trade
and investment to intellectual property limitations. I posit here that we
lack a holistic understanding of the nature, causes and implications of
regulatory shifts that suitably address transformative biotechnologies like
cell-cultivation in ever-expanding knowledge societies.
Future knowledge is contingent upon past forms thereof (Dupre and
Somsen 2019). However, whereas productivity has played an essential
role in traditional societies during the industrial age, knowledge is the
designated vector that generates social and economic results nowadays
(Drucker 1993). While this paradigm shift appeared long undisputed
(Cerroni 2020), it has been posited that the encoding of knowledge in
law precipitates a new knowledge capitalism that is particularly evident in
patent law as a fiction that limits cognitive processes with symbolic capi-
tal (Stehr 2022). Today, IPRs are closely linked to a productivist para-
digm that rewards efficient extraction of value-generating activity over
the creation of public welfare in its many facets. In this vein, the
commons are constitutive of a foundational element in capitalist intel-

lectual production (Broumas 2017), while potentially catapulting non-
rivalry in intellectual goods (Bracha 2018).
A flourishing innovation ecosystem is necessarily attuned to scientific
endeavours. Growing calls for open science as a form of co-creation may
find refuge in knowledge commons phenomena that emphasise their role
in scientific enquiry (Ranganathan 2016). Equally, large sets of informa-
tion (‘big data’) have generated significant options for open science and
the formation of knowledge societies (Wessels et al. 2017). As such, there
are alternatives to innovation incentives that purely rely on legal monop-
olies -as it is mostly the case with various regulatory approaches to IPRs.
For example, IP self-regulation has been considered as a basis for user
generated law in knowledge societies (Riis 2016), that is, a bottom-up
approach to regulation by IP.
With the aim of gaining a better understanding of ‘knowledge society’
epistemologies, I explore below the role of patents in cellular agriculture,
a field of enquiry that uses cell-cultivation technology, as a case study to
elucidate the extent to which IP rights can be deployed to generate opti-
mal public welfare while conserving and indeed expanding spheres of
intellectual commons, understood as the abstract embodiment of shared
resources available for all.
3 Patent Effectiveness and the WTO TRIPS

Arguably, the effectiveness of a patent is determined by its term of protec-
tion. Art. 33 TRIPS, in establishing a period of twenty years, significantly
boosts the strength of patent protection under the law of the World Trade
Organisation (WTO law). This is particularly true for fields of technol-
ogy where long-term research precedes the production of a marketable
invention, such cell-cultivation. Although today’s technology tends to
become obsolete more rapidly as the average period of useful life attrib-
uted to it by society decreases (Pires de Carvalho 2010), the average lifes-
pan of patents continues to correspond with the cycles of Schumpeter’s
“creative destruction” (Schumpeter 1947).
In setting the term of patent protection under TRIPS, Art. 33 does not
provide for any distinction on the basis of the field of technology, or
extent of exploitation, of patented inventions. Such distinctions would
arguably encourage discrimination among technological fields, leading to
an ad absurdum scenario whereby patent protection would be assigned
on a case-by-case basis, leading to extreme legal uncertainty (Pires de
Carvalho 2010).
Art. 33 stipulates that “[t]he term of protection available shall not end
before the expiration of a period of twenty years counted from the filing
date.” The term of a patent may not be shorter than 20 years, even if the
actual effects of the patent do not occur until after its grant. Accordingly,
Art. 33 identifies the earliest possible date for the end of patent protec-
tion. Although the provision ties the patent term to the date of filing and
does not contain specific regulations relating to the calculation of time
limits, this aspect was clarified by the Appellate Body in Canada— Patent
Term. In its view, the calculation of a term of patent protection is deter-
mined by taking the date of filing of the patent application and adding
twenty years (AB Canada – Patent Term, para. 85). Conversely, Members
are not bound to compensate for delays, for instance, in the examination
process of the application or in the marketing approval of products
(Correa 2022). As a result, the national patent regulations of Members
remain decisive. However, patent protection may not be significantly
curtailed by unreasonable delays in the procedures for granting patents,
as such delays are inconsistent with the obligations under Art. 62.1 and
2 TRIPS.
More specifically, the Panel in Canada—Patent Term found that Art.
33 establishes a minimum term of protection for patents, which also
raised questions about the possibility of extending rights and obligations
even after the expiry of said patent terms, which would generate spillover
effects (Canada – Patent Term, paras 6.57-6.121). The Panel Report was
subsequently appealed and, relying on the availability requirement in
Art. 33, the Appellate Body considered the provision “straightforward” in
stipulating that the filing date plus 20 years is the earliest date on which
the term of protection of a patent may end and that this 20-year term
must be “a readily discernible and specific right, and it must be clearly
seen as such by the patent applicant when a patent application is filed”
(AB Canada-Patent Term, paras 84 et seq.). On that basis, the Appellate

Body rejected Canada’s defensive argument, according to which other
statutory and regulatory provisions would allow patent applicants to
delay the procedure so as to extend the patent term to one de facto equiv-
alent to the term laid down in Art. 33 (AB Canada-Patent Term, paras 94
et seq.). It held that:
The opportunity to obtain a twenty-year patent term must not be ‘avail-

able’ only to those who are somehow able to meander successfully through
a maze of administrative procedures. The opportunity […] must be a read-
ily discernible and specific right, and it must be clearly seen as such by the
patent applicant when a patent application is filed (AB Canada – Patent
Term, para. 92).
Thus, in view of the Appellate Body’s understanding, Art. 33 does not

support an “equivalence test” whereby a combination of procedures,
when used in a particular sequence, may add up to twenty years.
Art. 33 does not, however, regulate further terms of protection existing
independently of a patent, such as supplementary protection certificates
(SPCs). The legitimacy of SPCs can be derived from Art. 33, which only
determines the earliest possible date for the end of patent protection. If
the patent itself may remain in force longer, this must hold true also for
SPCs after the expiry of the patent term. Art. 1.1, sentence 2 supports
this interpretation. However, Art. 33 only relates to the patent protection
itself. Accordingly, supplementary protection may not be deducted from
the lifetime of the patent. The protection by the patent itself shall not end
before the expiry of the time limit specified in Art. 33. This is also appli-
cable if the gradual interaction of the patent and supplementary protec-
tion operates such that the invention will be protected like a patent at
least until the expiry of such time. Art. 33 thus outlaws situations where
the protection of a patent itself would end before the expiry of the 20
years and then be substituted by supplementary protection. There is no
general requirement in TRIPS to conduct substantive examinations of
patent applications.16 However, it is Members’ obligation to put in place
a scrutinizing mechanism that assesses whether a claimed invention com-
plies with the mandatory conditions of patentability under Arts. 27.1
and 29.1 TRIPS. This requirement also applies to the re-registration of

foreign patents.
I have argued elsewhere (de Amstalden 2021) that while regulatory
transparency requirements can be transformative, it is not the only factor
with the potential to play a beneficial role in the shift from intensive
agricultural practices to more efficient, sustainable biotechnologies.
Patents are designed to promote invention and innovation by granting
patent holders a comprehensive competitive advantage against third party
imitation in return for public disclosure of information about the inven-
tion (Granstrand 2019). It is crucial to recognise that there are critical
public transparency implications associated with patents. Misconceptions
about IP rights, and patents in particular, have led the general public to
believe that patents hide information from third parties. Quite to the
contrary, a patent cannot be successfully filed without an accurate disclo-
sure of its inventiveness. Public disclosure of information is the price to
pay in return for a time-limited monopoly, essentially to recover research
and development investment. As such, the protection afforded by patents
is not confidentiality – as it is the case with trade secrets- but an exclusive
right to use the invention and license the technology if so desired.
Through this lens, market action relying on trade secrets as the main IP
strategy would indeed run contrary to increasingly robust calls for ‘open
science’ in cellular agriculture, for it is precisely the confidentiality
demanded by trade secrets that considerably hinders technological devel-
opment behind closed (and locked) doors.
Specifically for novel foods, patenting processes -as opposed to prod-
ucts- tend to be less stringent, as the patentability requirements of nov-
elty and inventiveness are easier to meet. Patents also afford the benefits
of awarding credibility, as well as the basis for licensing of technologies.
In fast-paced technological fields, as it is the case with cellular agriculture,
the term of protection conferred by patents is an attractive alternative
that warrants the (at times onerous) filing procedures. It is telling that the
last years saw a significant increase in the number of patent applications
in the cellular agriculture sector (Ng et al. 2021). Arguably, this develop-
ment may be explained via a number of reasons, which are potentially
two sides of the same coin. First, companies had no incentive to date to
file for patent applications earlier to conserve their technological
advantage over competitors through trade secrets law via enforceable

confidentiality agreements, as we have seen previously. Secondly, compa-
nies could have been incentivised by venture capitalists to file for patents
in order to successfully secure capital for expansion. Being mindful of the
divergent promissory narratives in cellular agriculture discourse (de
Amstalden 2021; Stephens et al. 2019), it is also conceivable that reluc-
tance to publicise the obtention of patents in the cellular agriculture sec-
tor is connected to the potential for consumer confusion and the
reinforcement of perceptions about the ‘unnaturalness’ of the novel foods.
4 Conclusion
This chapter explored the impact of transformative biotechnologies like
cell-cultivation on ever-expanding knowledge societies. In doing so, it
examined the role of patents, and their terms of protection under inter-
national IP law to elucidate whether and to what extent IPR as conceived
today are compatible with the calls for open science and the solidification
of an intellectual commons.
Just as cell-cultivation technology has the potential to be considered a
‘technology of abundance’, IP rights can, and indeed, should be con-
strued as an abundant resource. In doing so, it demonstrated, with pat-
ents as an example, that IP rights have the ability to invigorate multiple
tonalities in new global economic governance mechanisms, while being
mindful of, and in fact amplify, a variety of seemingly unrelated elements
uniting to address complex social challenges. Whether IP rights as public
interest mechanisms succeed in advancing policy considerations to rei-
magine food systems through cellular agriculture merits further empirical
research.
References
(Maidana-Eletti) de Amstalden, Mariela (2016), Global Food Governance:
Implications of Food Safety and Quality Standards in International Trade
Law, Peter Lang Publishers.
Bracha, Oren (2018), Give Us Back Our Tragedy: Nonrivalry in Intellectual

Property Law and Policy, 19 Theoretical Enquiries in Law, 633.
Broumas, Antonios (2017), The Ontology of the Intellectual Commons,
International Journal of Communication 11, 1507–1527
Cerroni, Andrea (2020), Understanding the Knowledge Society: A New
Paradigm in the Sociology of Knowledge, Edward Elgar 2020.
Correa, Carlos (2022), Interpreting the Flexibilities Under the TRIPS
Agreement, Springer.
De Amstalden, Mariela (2021), Seafood without the Sea, 23 Journal of World
Trade and Investment 1, 68–94.
Drucker, Peter F. (1993) “The rise of the knowledge society.” The Wilson
Quarterly, vol. 17, no. 2, spring 1993, pp. 52 et seq.
Dupré, S. and Somsen, G. (2019). The History of Knowledge and the Future of
Knowledge Societies. Berichte zur Wissenschaftsgeschichte, 42, 186–199.
https://doi.org/10.1002/bewi.201900006
Granstrand, O. (2019). Patents and policies for innovations and entrepreneur-
ship. In: Takenaka, T. (Ed), Research Handbook on Patent Law and Theory,
Edward Elgar, 55–86.
Neil Stephens, Alexandra Sexton, Clemens Driessen, ‘Making Sense of Making
Meat: Key Moments in the First 20 Years of Tissue Engineering Muscle to
Make Food’, Frontiers in Sustainable Food Systems, Volume 3 - 2019.
https://doi.org/10.3389/fsufs.2019.00045
Ng, E., Singh, S., Yap, W., Tay, S., Choudhury, D., 2021. Cultured meat: a
patentometric analysis. Critical Reviews in Food Science and Nutrition,
https://doi.org/10.1080/10408398.2021.1980760.
Pires de Carvalho, Nunes (2010), The Trips Regime of Patent Rights, 3rd
Edition, Wolters Kluwer.
Ranganathan, Surabhi (2016), Global Commons, European Journal of
International Law, 27(3), 693–717.
Riis, Thomas (2016), User Generated Law: Re-Constructing Intellectual
Property Knowledge in the Knowledge Society, Edward Elgar.
Schumpeter, Joseph (1947), The Creative Response in Economic History, 7 The
Journal of Economic History 2, November 1947, pp. 149 – 159.
Stehr, Nico (2022), Knowledge Capitalism, Routledge.
Wessels, Bridgette, Finn, Rachel, Sveinsdottir, Thordis and Wadhwa, Kush
(2017) Open Data and the Knowledge Society’, Amsterdam University
Winston Churchill, ‘Fifty Years Hence’, The Strand Magazine (December 1931).
Annexure: Table of Cases
CLM v CLN [2022] SGHC 46

Coca-Cola v Barr
Countess Housewares v Addis
Infopaq International A/S v Danske Dagbaldes Forening C- 5/08
IRC v Muller & Co’s Margarine Ltd [1901]
Janesh s/o Rajkumar v Unknown Person [2022]
National Provincial Bank Ltd v Ainsworth [1965]
Osbourne v Persons Unknown [2022]
Ruscoe v Cryptopia Ltd (in liq) [2020]
Sheraton Corp. of America v Sheraton Motels [1964]
Strix Ltd v Otter Controls Ltd [1995]
Thaler v Comptroller General of Patents Trade Marks And Designs
[2021] EWCA Civ 1374
Thaler v Commissioner of Patents [2021] FCA 879
Visual Entidad De Gestion De Artistas Plasticos/ Punto Fa, S.L. [2022]
AJM B 1900/2022 – ECLI:ES: JMB: 2022:1900
https://doi.org/10.1007/978-3-031-42576-9
Index
A B
Ageing population, xi, 129–133 Berne Convention, 94, 158
Algorithms, 2, 5, 13, 30, 71, 107, Big data, vi, x, 5, 25–60, 100, 114,
108, 115–119, 124, 125 143, 184
Artificial Intelligence (AI), v–vii, Biotechnologies, vi, xii, 57, 181–188
ix–xii, 1–23, 26, 30, 40, 71, Blockchain, xi, 65–70, 72, 75, 76,
91–108, 113–115, 117–125, 79, 82, 84
130, 134–137, 139, 141, Blockchain technology, 65–67, 70,
176, 178 71, 75, 76, 84
assisted, 2, 4–6, 9–11, 15, 16, 21,
22, 176, 178
generated, 2, 4, 6, 9–13, 15, 16, C
19, 21, 22, 93–96, 135, Canada—Patent Term, 185, 186
176, 178 Care robots, xi, 129–152
ownership, vi, 1–23 Cell-cultivation technology, xii,
supported, 4, 5, 21, 22 181–184, 188
technology, v, ix, x, 5–7, 26, 40, 99, Cellular agriculture, xii, 181–184,
114, 115, 117–121, 124–126 187, 188
Artwork, vi, xi, 13, 65–85, 97, 108 Centre for International Intellectual
Assignment, 168, 175 Property Studies (CEIPI), 12
https://doi.org/10.1007/978-3-031-42576-9
194 Index
Climate change, xii, 115, 182 Data mining, 30, 31, 97–99
Compendium of US Copyright Data ownership, 31–35, 45
Practices, 14 Data security, 131, 143–145, 149
Computer generated works (CGW), Data transfers, 37, 42, 45, 47,
10, 16, 19, 20, 22, 23, 95, 108 48, 51
Confidentiality, 39, 47, 142, 160, Deception, xi, 132, 134–137, 141
161, 167, 187, 188 Derivative data, 44
Consumer protection, 45–47, 149, Design Rights Audit, 162–163
150, 183 Digital economy, 27, 28, 30, 31,
Contextual adaptation, 16 35–40, 51, 52, 57, 58, 60
Contractual clauses, 6, 21, 167 Digital Markets Act, 39, 48
Copyright, xi, xii, 2, 6, 9–12, 14–16, Due diligence, 168–169, 174, 177
19–20, 22, 67, 70–73, 80–83,
91–101, 105–108, 115, 116,
122, 157, 160, 163, 167, 168, E
174, 177 Emerging technologies, ix, 1, 176
Copyright Audit, 167–168 Ethereum, 66, 68, 69, 75
Copyright Design Patent Act Ethical business strategy, 17
(CDPA) 1988, 9, 10, 19, Ethical discourse, 17
95, 98, 101 Ethics, ix, 6, 17, 20, 23, 124, 125,
Cost method, 170–172 132, 149, 176
Crypto assets, 73–75, 77 European Court of Justice (CJEU),
Cryptocurrency, xi, 66, 67, 69, 12, 77, 93, 94
73–75, 77, 78, 80, 83 European Patent Convention
Culpability, 20 (EPC), 8
Cyber law, 122 European Patent Office (EPO),
Cyberspace, 114–116, 122 8, 18, 103
European Union Commission, 5–7,
11, 22, 26
D European Union Commission IP
DABUS, xi, 8–9, 22, 101–107 Action Plan, 11
DARPA, 3 European Union Trade Marks
Databases, 5, 18, 20, 33, 42, 79, 97, Directive, 163
99, 162, 167
Database Directive, 33, 41, 42, 51
Data enclosures, 26, 27, 41, 43, F
50, 51, 58 Fair, reasonable and non-
Data Governance Act, 35, 39, discriminatory (FRAND)
40, 52, 60 terms, 48, 55
Index 195
Fourth industrial revolution (4IR), M

18, 27–38, 44 Madrid Goods and Service
Franchising, 161, 169, 175, 178 (MGS), 18
Fungible tokens, 69 Manipulation, 134–137, 141
Market value approach, 173–174
Metadata, 72, 76, 81, 83, 96
G Minting, 72, 80–82
General Data Protection Regulation
(GDPR), 29, 34, 39, 40, 43,
45, 46, 51, 142, 144 N
Generated data, 29, 33, 45, 47, 48 Natural Language Processing
Goodwill, x, 165–168, 170 (NLP), 18
Greenhouse gas (GHG) The Next Rembrandt, 2
emissions, 182 Non-fungible tokens (NFTs), vi, x,
xi, 66, 67, 69–84
Non-human actors, 14–15
H
Hidden assets, 167, 170
Human food consumption, 181–183 P
Passing off, 165–167
Patent Act 1977, 162
I Patent audit, 162
Idea expression dichotomy, 9, 10 Patent Cooperation Treaty
Income-based approach, 172 (PCT), 103
Indonesian Patent Law, xi, 117, 118, Patent process, 119, 162
121, 126 Patent protection, 7, 9, 55, 56, 107,
Institute for Information Law 116, 120, 184–186
(IViR), 12 Patent registration, 7, 103, 162
Intellectual commons, xii, 181–188 Patent term, 185, 186
Intellectual property audit, x Pathetic Dot theory, 122, 123, 125
Intellectual property valuation, Privacy, 32, 34–39, 44, 132, 133,
161, 170 142–145, 148–151
International Association for the Productivist paradigm, 183
Protection of Intellectual
Property (AIPPI), 94, 107
R
Regulation on Harmonised Rules on
K Fair Access to and Use of Data
Know-how, x, xii, 157, 167, 177 (Data Act), vi, 25–60
196 Index
Robotics, v, ix, xi, 1, 130, 133, 145, Transformative biotechnologies, xii,

146, 148–150 181–183, 188
Royalties, 67, 173–175 Trustworthy AI, 17, 20, 135
Turbo Tax, 3, 21
S
Small and medium-sized enterprises U
(SMEs), 34, 48, 100 UK Intellectual Property Office
Smart contracts, 47, 65, 66, 70, 74, (UKIPO), 101–103, 164
75, 82, 83 User generated data, 26, 27, 29, 32,
Smart technologies, x, 25–27, 29, 30 44, 45, 50
Source code, 119, 124–125 User safety, xi, 144–150
Sui generis right, 10, 33, 41, 42, 51 US Patent and Trademark Office
Supplementary protection certificates (USPTO), 18, 103, 104,
(SPC), 186 120, 121
T V
Text and data mining (TDM), Visual Entidad de Gestión de
20, 97–101 Artistas Plásticos
Trademark, xii, 19, 57, 78–80, 157, (VEGAP), 80, 81
159, 161, 163–165, 177, 183
Trade Mark International
Classification Services W
(TMICS), 18 World Health Organisation
Trademarks audit, 163–165 (WHO), 129
Trade Related Intellectual Property World Intellectual Property Office
Rights (TRIPS), 7–9, 17, 116, (WIPO), 2–4, 7, 18, 21, 92,
118, 125, 126, 162, 184–188 117, 158, 173
Trade secrets, 33, 36, 46, 47, 57, 58, World Trade Organisation
167, 177, 183, 187, 188 (WTO), 184

Developments in Intellectual Property Strategy: The Impact of Artificial Intelligence, Robotics and New Technologies

Uploaded by

Copyright:

Available Formats

Developments in Intellectual Property Strategy: The Impact of Artificial Intelligence, Robotics and New Technologies

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Developments in Intellectual Property Strategy: The Impact of Artificial Intelligence, Robotics and New Technologies

Uploaded by

Copyright:

Available Formats

Edited by

ISBN 978-3-031-42575-2 ISBN 978-3-031-42576-9 (eBook)

Paper in this product is recyclable.

The contemporary debates that currently rage in academe and social

review of the rapidly evolving fields of artificial intelligence and conse-

Emeritus Professor International Law David Price

Developments in intellectual property strategy is a cutting-edge and

technologies. It considers different legal and regulatory governance

discussion against the background of the trade of artworks in a block-

determining the appropriate levels of regulatory oversight and support

Birmingham, UK Nadia Naim

 Artificial Intelligence Invention Protection Model113

3 Conclusion & Suggestion 125

 Examination of the Tangible Value of IP Financing for

4 IP Due Diligence 168

Abdurrahman Alfaqiih is a lecturer in the Department of Civil Law also the

Department of Philosophy and a research assistant at the Centre for Biomedical

As part of his PhD research project, he worked at the German Collective

Angelia Jia Wang is an assistant professor in the Law School at Durham

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 1

intellectual property and artificial intelligence, intellectual property rights

the UK Intellectual Property Office (UK IPO) leading several consulta-

1 What Is Artificial Intelligence (AI)?

characteristics consistent with cognitive intelligence, it is not conscious,

2 Defining Artificial Intelligence Within

The second-wave of AI systems, which can be clearly seen as AI- assisted

3 EU Commission and Artificial Intelligence

4 So Where Does That Leave Us?

(1) The invention is new

Article 27 stated the requirements for; novelty, inventiveness, industrial

5 To DABUS or Not to DABUS

For patent protection, on the whole, as it currently stands the answer

6 The Copyright and AI Battle

Fig. 1 The idea expression dichotomy

There are several stages identified when assessing AI assistance in copy-

(1) a “production in the literary, scientific or artistic domain”;

Artificial intelligence in the music industry is a topic of interest, given

7 Can the Four-Step Test Be Applied

Despite the requirement for human intellectual creativity, the chal-

J. Walter Thompson Amsterdam (2016)

8 Non-human Actors in the Originality Test

input or intervention from a human author. The crucial question is

9 Artificial Intelligence: A Black Swan

AI-assisted and AI-generated waves, there needs to be an IP system in

permanently closed. Rather, as the next wave of AI develops to mimic the

10 Legal and Ethical Considerations

socio-­political conditions. Notwithstanding the need for specific criteria,

12 Patent Offices and AI

services (UK IPO Consultation, 2023). Legislation could be altered to

13 AI in Copyright Law

and methods, the legislative changes require unbiased and independent

14 AI in Patent Law

15 Alas, Who Should the Intellectual

Therefore, AI-supported systems are predictable by humans and there

Infopaq International A/S v Danske Dagblades Forening (C-5/08)

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 25

commonly referred to as Big Data. Thus, the collecting and processing of

2 Utilising the Value of Big Data:

Irion 2021). This digital economy can also be described as data-­driven in

Birmingham, UK Nadia Naim

Artificial Intelligence Invention Protection Model113

3 Conclusion & Suggestion 125

Examination of the Tangible Value of IP Financing for

4 IP Due Diligence 168

1 What Is Artificial Intelligence (AI)?

2 Defining Artificial Intelligence Within

3 EU Commission and Artificial Intelligence

4 So Where Does That Leave Us?

5 To DABUS or Not to DABUS

6 The Copyright and AI Battle

7 Can the Four-Step Test Be Applied

8 Non-human Actors in the Originality Test

9 Artificial Intelligence: A Black Swan

10 Legal and Ethical Considerations

socio-political conditions. Notwithstanding the need for specific criteria,

12 Patent Offices and AI

13 AI in Copyright Law

14 AI in Patent Law

15 Alas, Who Should the Intellectual

2 Utilising the Value of Big Data:

Irion 2021). This digital economy can also be described as data-driven in

2.2 The Importance of Big Data for Future

2.3 Unlocking the Value/Potential of Big Data

2.3.2 A Right to Access

3 The EU Data Strategy and the Data Act

3.2 The Data Act

3.2.1 B2B and B2C Data Access

3.2.2 B2G Data Access

(i) a public sector body, the Commission, the European Central

3.2.3 International Data Transfers, Interoperability

3.2.4 Initial Critique of the Data Act

4 A de-Facto Exclusive Right to Data

4.1 The Problem with Exclusivity: Lessons

4.2 An Exclusive Right to Data and the Problem

2 Blockchain and the Tokenisation

2.2 The Tokens: Fungible and Non-fungible

An NFT is a piece of digital artefact that represents the ownership of real-

3 The Trading of Artworks

4 Legal Status of NFT: The Common

4.1 The United Kingdom

5 Legal Status of NFT: The Civil

5.2 European Cases

5.3 Chinese Cases