Ray Campbell is a scholar of civil procedure and the legal profession. His current research concentrates on the intersection of law and commerce, with a special focus on the changing nature of the legal services marketplace in light of technological and economic innovations. He also writes on civil procedure topics, especially those with a transnational twist, and is creating a Civil Procedure textbook on the H2O platform specialized for non US students. Before embarking on his career in academia, Professor Campbell served as law clerk to U.S. Chief Justice Warren Burger of the United States Supreme Court, as a law clerk to Judge Malcolm Wilkey of the U.S. Court of Appeals for the D.C. Circuit, as a prominent litigation partner in the multinational law firms of Kirkland Address: Shenzhen, China
Artificial intelligence in judicial decision-making has gone from science fiction to reality... more Artificial intelligence in judicial decision-making has gone from science fiction to reality. While rulings from ‘robot judges’ in important cases remain the realm of futurists, AI already plays a critical role in courts around the world. AI’s increasing influence demands a nuanced approach. The uses for AI in chambers today include almost everything but the full delegation of significant cases. Overworked judges have admitted to using AI to be sure that their sense of normal outcomes is correct. Beyond ad hoc use, in China data analysis of the facts and law in a case allows judges and their supervisors to compare outcomes to predicted outcomes based on the dataset; this guides judges and also helps supervisors identify divergent results that might be due to incompetence or corruption. Many courts use AI to help guide the public as they encounter an unfamiliar legal system, whether that takes the form of robots greeting them at the courthouse entrance or guiding them step by step through online case submission. Transcript creation, document and transcript summaries, document review, AI-aided legal research, scheduling hearings, and translating documents are also among AI’s uses. More controversially, AI allows algorithmic recommendations as to whether pre-trial release or probation should be granted to criminals and what level of punishment should be imposed. The appetite for AI adoption varies by jurisdiction. In his most recent annual message on the U.S. judiciary, Chief Justice Roberts praised the use of technology but also made clear that human decision-making was preferred. In China, by contrast, courts are required to implement a wide range of artificial intelligence tools by 2025. England and Wales has led in allowing non-lawyers, including AI-powered non-lawyers, to offer services to the public, while Singapore has embraced technology as a way to enhance its competitive appeal based on the rule of law. This increased use of artificial intelligence in deciding cases raises both practical and philosophical questions. Biases are captured in data that reflect past, often discriminatory, practices, and these can infect AI recommendations. Due process concerns arise when any part of the right to a hearing is delegated to an alien, non-human intelligence or when opaque, ‘black-box’ AI tools obscure the basis for recommendations and decisions. Even when the use of AI is disclosed and any final decision rendered by humans, the anchoring effect and undue deference to technology can change outcomes. Beyond all the practical issues is whether humans want important issues decided for them by machines. This article concludes by looking at where advances in AI capabilities might take courts in the future. As technology advances and user resistance drops, AI will be available for more tasks. Using AI appropriately will require a nuanced, technically sophisticated approach. AI influence on the process will not be cured by ‘judge washing’ the algorithm with a final human decision maker. Just as ethical rules control and limit the influence other humans can have on judicial decisions, systems and rules will have to evolve for AI to make sure the influence of algorithms is appropriately controlled.
A Jotwell review of Richard Susskind and Daniel Susskind, The Future of the Professions (2016) Th... more A Jotwell review of Richard Susskind and Daniel Susskind, The Future of the Professions (2016) This brief review examines the thesis advanced by Richard and Daniel Susskind in the Future of the Professions - that advancing technology will allow us to manage information, including expert information and know how of the kind traditionally associated with professions, in a way that makes professionals unnecessary.
ABSTRACT Parked at the crossroads of higher education and legal practice, U.S. law schools find t... more ABSTRACT Parked at the crossroads of higher education and legal practice, U.S. law schools find their once serene setting to be under siege. The legal services industry faces a period of profound structural change with uncertain consequences for the traditional practice of law. Meanwhile, higher education wades through its own Slough of Despond, with its mission unclear and elite institutions spooked by disruptive change.Just at this moment Brian Tamanaha has entered the scene with a book entitled Failing Law Schools. Tamanaha argues persuasively that law schools are in crisis, with costs so high and employment prospects so poor that most law schools now represent a bad investment for students. He tells how things came to this sad pass, with legal educator selfishness playing the leading role. He builds a compelling case. Tamanaha falters when he attempts to present a path forward for law schools. In proposing changes for the future, he essentially seeks a return to a more affordable past. In the context of changes sweeping higher education and legal services, such a return may not be possible or desirable. Law schools that find themselves in a challenging situation will need to address – and perhaps embrace – radical change in order to chart a path forward.Tamanaha’s failure to address the future flows from fundamental limitations in his approach. First, Tamanaha errs in treating law schools as a special case. The most severe ills afflicting law schools – the orientation towards benefiting faculty rather than serving students, a system-wide elevation of research over teaching, and soaring tuitions – appear in similar degree across the academic spectrum of modern U.S. universities. Second, Tamanaha also fails to engage with some of the most disruptive (and promising) changes coming to higher education, including eventually law schools. Effective online learning threatens to split the instruction aspect of higher education from the research and networking aspects, with enormous implications for educational finances as well as institutional missions. Tamanaha’s modest “back to the future” vision fails to engage with the ways these new developments create opportunities for schools to be better than they ever have been, at lower costs.Tamanaha also fails to take into account the changing nature of the legal services market. He focuses on training traditional lawyers for traditional legal practice. Both at the high and low ends, however, change is coming to the legal services market, impacting both the demand for traditionally schooled lawyers and potentially creating demand for new kinds of service providers. Tamanaha’s prescription fails to address this systemic change, and as a result fails to look at how law schools might be best reconfigured to serve a society with changing needs.That said, Tamanaha has helped launch a discussion that needs to be had. No professor with a conscience can comfortably watch half his or her students spend three years of their youth and significant sums on a legal education only to find no jobs that justify the investment. For those schools not in the very upper tier of American legal education, something needs to change, and Tamanaha’s book helps structure the discussion of what that change should be.
In recent years an increasing number of students who are not likely to practice in U.S. courts ha... more In recent years an increasing number of students who are not likely to practice in U.S. courts have been enrolled in U.S. civil procedure courses. While their needs differ from domestic students, those practicing even abroad in a world of legal pluralism and polycentric governance will be handicapped if they do not understand the legal process of the world's largest economy. Based on 13 years of teaching non-U.S. students at the Peking University School of Transnational Law, this article summarized my experience and observations. Students today need to understand how polycentric governance makes legal systems beyond their own important. They also, in some cases, need more depth on issues such as cross-border discovery or enforcement of foreign judgments in the U.S. than the average introductory U.S. civil procedure course is likely to give. In order to provide what I thought my students needed, I found it necessary to develop a textbook, which is available without charge at https://opencasebook.org/casebooks/7581-us-civil-procedure-for-international-students-2022-2023-edition/
In recent years an increasing number of students who are not likely to practice in U.S. courts ha... more In recent years an increasing number of students who are not likely to practice in U.S. courts have been enrolled in U.S. civil procedure courses. While their needs differ from domestic students, those practicing even abroad in a world of legal pluralism and polycentric governance will be handicapped if they do not understand the legal process of the world's largest economy. Based on 13 years of teaching non-U.S. students at the Peking University School of Transnational Law, this article summarized my experience and observations. Students today need to understand how polycentric governance makes legal systems beyond their own important. They also, in some cases, need more depth on issues such as cross-border discovery or enforcement of foreign judgments in the U.S. than the average introductory U.S. civil procedure course is likely to give. In order to provide what I thought my students needed, I found it necessary to develop a textbook, which is available without charge at https://opencasebook.org/casebooks/7581-us-civil-procedure-for-international-students-2022-2023-edition/
Ray Worthy Campbell, Artificial Intelligence in the Courtroom: The Delivery of Justice in the Age of Machine Learning, 18 COLO. TECH. L.J. 323 (2020) , 2020
How realistic is the idea that AI can take over from human judges? Can we expect AI software to t... more How realistic is the idea that AI can take over from human judges? Can we expect AI software to take over the role that judges play anytime in the near future? A realistic view of what role AI can play in the judicial process requires first a look at how AI operates and what functions it can be expected to perform in the near to intermediate term. In addition, in light of the developing field of judicial studies, we must look--as other commentators in this area have not--at the full scope of the judicial function, which goes far beyond just resolving individual cases. Only by matching the real potential of AI with the full range of judicial functions can we give a non-hyperbolic assessment. Setting aside the possibility of radical technological advances, what we can expect in the near term is for software to play a role supporting--but not replacing--human judges. AI can, in certain cases, predict how a certain case might come out. That, however, falls far short of what judges do. The current capability of AI is limited to specialized tasks, and the roles of judges are so generalized that there is no near-term possibility of AI wholly and satisfactorily displacing judges in high stakes cases. Even if software is developed to perform generalist capabilities, an uncertain prospect at best,4 we must still face the issue of whether we are prepared to delegate the creation and application of legal rights and responsibilities to impersonal artificial entities. In this regard it is important to remember that even if AI makes *325 the leap to general intelligence, such AI will in no way be human. Put simply, while in some ways as capable as human beings, AI will remain alien in fundamental ways. Whether such an intelligence can fulfill the diverse and fundamental roles played by human judges requires value choices beyond the scope of technology.
In recent years, substantial friction has been generated by United States litigation involving Ch... more In recent years, substantial friction has been generated by United States litigation involving Chinese defendants. From the U.S. perspective, the litigation is primarily domestic in nature, seeking compensation for injuries caused on U.S. soil by parties that actively targeted the U.S. market. Aside from obtaining compensation for those injured, a purpose of the litigation is to create deterrence so as to discourage future violations. Obtaining discovery, which is critical to U.S. style litigation, is an essential part of that ‘after-the-fact regulation’ process.From the Chinese perspective, foreign court systems imposing demands on Chinese defendants that need to be carried out on Chinese soil constitute an affront to Chinese sovereignty. Familiar with a system in which discovery is not at the core of the dispute resolution process, and in which litigant-driven discovery plays almost no part at all, they see no justification for U.S. courts requiring compliance with U.S. style disc...
A Jotwell review of Richard Susskind and Daniel Susskind, The Future of the Professions (2016) Th... more A Jotwell review of Richard Susskind and Daniel Susskind, The Future of the Professions (2016) This brief review examines the thesis advanced by Richard and Daniel Susskind in the Future of the Professions - that advancing technology will allow us to manage information, including expert information and know how of the kind traditionally associated with professions, in a way that makes professionals unnecessary.
Artificial intelligence in judicial decision-making has gone from science fiction to reality... more Artificial intelligence in judicial decision-making has gone from science fiction to reality. While rulings from ‘robot judges’ in important cases remain the realm of futurists, AI already plays a critical role in courts around the world. AI’s increasing influence demands a nuanced approach. The uses for AI in chambers today include almost everything but the full delegation of significant cases. Overworked judges have admitted to using AI to be sure that their sense of normal outcomes is correct. Beyond ad hoc use, in China data analysis of the facts and law in a case allows judges and their supervisors to compare outcomes to predicted outcomes based on the dataset; this guides judges and also helps supervisors identify divergent results that might be due to incompetence or corruption. Many courts use AI to help guide the public as they encounter an unfamiliar legal system, whether that takes the form of robots greeting them at the courthouse entrance or guiding them step by step through online case submission. Transcript creation, document and transcript summaries, document review, AI-aided legal research, scheduling hearings, and translating documents are also among AI’s uses. More controversially, AI allows algorithmic recommendations as to whether pre-trial release or probation should be granted to criminals and what level of punishment should be imposed. The appetite for AI adoption varies by jurisdiction. In his most recent annual message on the U.S. judiciary, Chief Justice Roberts praised the use of technology but also made clear that human decision-making was preferred. In China, by contrast, courts are required to implement a wide range of artificial intelligence tools by 2025. England and Wales has led in allowing non-lawyers, including AI-powered non-lawyers, to offer services to the public, while Singapore has embraced technology as a way to enhance its competitive appeal based on the rule of law. This increased use of artificial intelligence in deciding cases raises both practical and philosophical questions. Biases are captured in data that reflect past, often discriminatory, practices, and these can infect AI recommendations. Due process concerns arise when any part of the right to a hearing is delegated to an alien, non-human intelligence or when opaque, ‘black-box’ AI tools obscure the basis for recommendations and decisions. Even when the use of AI is disclosed and any final decision rendered by humans, the anchoring effect and undue deference to technology can change outcomes. Beyond all the practical issues is whether humans want important issues decided for them by machines. This article concludes by looking at where advances in AI capabilities might take courts in the future. As technology advances and user resistance drops, AI will be available for more tasks. Using AI appropriately will require a nuanced, technically sophisticated approach. AI influence on the process will not be cured by ‘judge washing’ the algorithm with a final human decision maker. Just as ethical rules control and limit the influence other humans can have on judicial decisions, systems and rules will have to evolve for AI to make sure the influence of algorithms is appropriately controlled.
A Jotwell review of Richard Susskind and Daniel Susskind, The Future of the Professions (2016) Th... more A Jotwell review of Richard Susskind and Daniel Susskind, The Future of the Professions (2016) This brief review examines the thesis advanced by Richard and Daniel Susskind in the Future of the Professions - that advancing technology will allow us to manage information, including expert information and know how of the kind traditionally associated with professions, in a way that makes professionals unnecessary.
ABSTRACT Parked at the crossroads of higher education and legal practice, U.S. law schools find t... more ABSTRACT Parked at the crossroads of higher education and legal practice, U.S. law schools find their once serene setting to be under siege. The legal services industry faces a period of profound structural change with uncertain consequences for the traditional practice of law. Meanwhile, higher education wades through its own Slough of Despond, with its mission unclear and elite institutions spooked by disruptive change.Just at this moment Brian Tamanaha has entered the scene with a book entitled Failing Law Schools. Tamanaha argues persuasively that law schools are in crisis, with costs so high and employment prospects so poor that most law schools now represent a bad investment for students. He tells how things came to this sad pass, with legal educator selfishness playing the leading role. He builds a compelling case. Tamanaha falters when he attempts to present a path forward for law schools. In proposing changes for the future, he essentially seeks a return to a more affordable past. In the context of changes sweeping higher education and legal services, such a return may not be possible or desirable. Law schools that find themselves in a challenging situation will need to address – and perhaps embrace – radical change in order to chart a path forward.Tamanaha’s failure to address the future flows from fundamental limitations in his approach. First, Tamanaha errs in treating law schools as a special case. The most severe ills afflicting law schools – the orientation towards benefiting faculty rather than serving students, a system-wide elevation of research over teaching, and soaring tuitions – appear in similar degree across the academic spectrum of modern U.S. universities. Second, Tamanaha also fails to engage with some of the most disruptive (and promising) changes coming to higher education, including eventually law schools. Effective online learning threatens to split the instruction aspect of higher education from the research and networking aspects, with enormous implications for educational finances as well as institutional missions. Tamanaha’s modest “back to the future” vision fails to engage with the ways these new developments create opportunities for schools to be better than they ever have been, at lower costs.Tamanaha also fails to take into account the changing nature of the legal services market. He focuses on training traditional lawyers for traditional legal practice. Both at the high and low ends, however, change is coming to the legal services market, impacting both the demand for traditionally schooled lawyers and potentially creating demand for new kinds of service providers. Tamanaha’s prescription fails to address this systemic change, and as a result fails to look at how law schools might be best reconfigured to serve a society with changing needs.That said, Tamanaha has helped launch a discussion that needs to be had. No professor with a conscience can comfortably watch half his or her students spend three years of their youth and significant sums on a legal education only to find no jobs that justify the investment. For those schools not in the very upper tier of American legal education, something needs to change, and Tamanaha’s book helps structure the discussion of what that change should be.
In recent years an increasing number of students who are not likely to practice in U.S. courts ha... more In recent years an increasing number of students who are not likely to practice in U.S. courts have been enrolled in U.S. civil procedure courses. While their needs differ from domestic students, those practicing even abroad in a world of legal pluralism and polycentric governance will be handicapped if they do not understand the legal process of the world's largest economy. Based on 13 years of teaching non-U.S. students at the Peking University School of Transnational Law, this article summarized my experience and observations. Students today need to understand how polycentric governance makes legal systems beyond their own important. They also, in some cases, need more depth on issues such as cross-border discovery or enforcement of foreign judgments in the U.S. than the average introductory U.S. civil procedure course is likely to give. In order to provide what I thought my students needed, I found it necessary to develop a textbook, which is available without charge at https://opencasebook.org/casebooks/7581-us-civil-procedure-for-international-students-2022-2023-edition/
In recent years an increasing number of students who are not likely to practice in U.S. courts ha... more In recent years an increasing number of students who are not likely to practice in U.S. courts have been enrolled in U.S. civil procedure courses. While their needs differ from domestic students, those practicing even abroad in a world of legal pluralism and polycentric governance will be handicapped if they do not understand the legal process of the world's largest economy. Based on 13 years of teaching non-U.S. students at the Peking University School of Transnational Law, this article summarized my experience and observations. Students today need to understand how polycentric governance makes legal systems beyond their own important. They also, in some cases, need more depth on issues such as cross-border discovery or enforcement of foreign judgments in the U.S. than the average introductory U.S. civil procedure course is likely to give. In order to provide what I thought my students needed, I found it necessary to develop a textbook, which is available without charge at https://opencasebook.org/casebooks/7581-us-civil-procedure-for-international-students-2022-2023-edition/
Ray Worthy Campbell, Artificial Intelligence in the Courtroom: The Delivery of Justice in the Age of Machine Learning, 18 COLO. TECH. L.J. 323 (2020) , 2020
How realistic is the idea that AI can take over from human judges? Can we expect AI software to t... more How realistic is the idea that AI can take over from human judges? Can we expect AI software to take over the role that judges play anytime in the near future? A realistic view of what role AI can play in the judicial process requires first a look at how AI operates and what functions it can be expected to perform in the near to intermediate term. In addition, in light of the developing field of judicial studies, we must look--as other commentators in this area have not--at the full scope of the judicial function, which goes far beyond just resolving individual cases. Only by matching the real potential of AI with the full range of judicial functions can we give a non-hyperbolic assessment. Setting aside the possibility of radical technological advances, what we can expect in the near term is for software to play a role supporting--but not replacing--human judges. AI can, in certain cases, predict how a certain case might come out. That, however, falls far short of what judges do. The current capability of AI is limited to specialized tasks, and the roles of judges are so generalized that there is no near-term possibility of AI wholly and satisfactorily displacing judges in high stakes cases. Even if software is developed to perform generalist capabilities, an uncertain prospect at best,4 we must still face the issue of whether we are prepared to delegate the creation and application of legal rights and responsibilities to impersonal artificial entities. In this regard it is important to remember that even if AI makes *325 the leap to general intelligence, such AI will in no way be human. Put simply, while in some ways as capable as human beings, AI will remain alien in fundamental ways. Whether such an intelligence can fulfill the diverse and fundamental roles played by human judges requires value choices beyond the scope of technology.
In recent years, substantial friction has been generated by United States litigation involving Ch... more In recent years, substantial friction has been generated by United States litigation involving Chinese defendants. From the U.S. perspective, the litigation is primarily domestic in nature, seeking compensation for injuries caused on U.S. soil by parties that actively targeted the U.S. market. Aside from obtaining compensation for those injured, a purpose of the litigation is to create deterrence so as to discourage future violations. Obtaining discovery, which is critical to U.S. style litigation, is an essential part of that ‘after-the-fact regulation’ process.From the Chinese perspective, foreign court systems imposing demands on Chinese defendants that need to be carried out on Chinese soil constitute an affront to Chinese sovereignty. Familiar with a system in which discovery is not at the core of the dispute resolution process, and in which litigant-driven discovery plays almost no part at all, they see no justification for U.S. courts requiring compliance with U.S. style disc...
A Jotwell review of Richard Susskind and Daniel Susskind, The Future of the Professions (2016) Th... more A Jotwell review of Richard Susskind and Daniel Susskind, The Future of the Professions (2016) This brief review examines the thesis advanced by Richard and Daniel Susskind in the Future of the Professions - that advancing technology will allow us to manage information, including expert information and know how of the kind traditionally associated with professions, in a way that makes professionals unnecessary.
Teaching U.S. Civil Procedure to Non-U.S. Students: Educating Students for a World of Legal Pluralism, 2023
In recent years an increasing number of students who are not likely to practice in U.S. courts ha... more In recent years an increasing number of students who are not likely to practice in U.S. courts have been enrolled in U.S. civil procedure courses. While their needs differ from domestic students, those practicing even abroad in a world of legal pluralism and polycentric governance will be handicapped if they do not understand the legal process of the world's largest economy. Based on 13 years of teaching non-U.S. students at the Peking University School of Transnational Law, this article summarized my experience and observations. Students today need to understand how polycentric governance makes legal systems beyond their own important. They also, in some cases, need more depth on issues such as cross-border discovery or enforcement of foreign judgments in the U.S. than the average introductory U.S. civil procedure course is likely to give. In order to provide what I thought my students needed, I found it necessary to develop a textbook, which is available without charge at https://opencasebook.org/casebooks/7581-us-civil-procedure-for-international-students-2022-2023-edition/
Uploads
Papers by Ray Campbell
The uses for AI in chambers today include almost everything but the full delegation of significant cases. Overworked judges have admitted to using AI to be sure that their sense of normal outcomes is correct. Beyond ad hoc use, in China data analysis of the facts and law in a case allows judges and their supervisors to compare outcomes to predicted outcomes based on the dataset; this guides judges and also helps supervisors identify divergent results that might be due to incompetence or corruption. Many courts use AI to help guide the public as they encounter an unfamiliar legal system, whether that takes the form of robots greeting them at the courthouse entrance or guiding them step by step through online case submission. Transcript creation, document and transcript summaries, document review, AI-aided legal research, scheduling hearings, and translating documents are also among AI’s uses. More controversially, AI allows algorithmic recommendations as to whether pre-trial release or probation should be granted to criminals and what level of punishment should be imposed.
The appetite for AI adoption varies by jurisdiction. In his most recent annual message on the U.S. judiciary, Chief Justice Roberts praised the use of technology but also made clear that human decision-making was preferred. In China, by contrast, courts are required to implement a wide range of artificial intelligence tools by 2025. England and Wales has led in allowing non-lawyers, including AI-powered non-lawyers, to offer services to the public, while Singapore has embraced technology as a way to enhance its competitive appeal based on the rule of law.
This increased use of artificial intelligence in deciding cases raises both practical and philosophical questions. Biases are captured in data that reflect past, often discriminatory, practices, and these can infect AI recommendations. Due process concerns arise when any part of the right to a hearing is delegated to an alien, non-human intelligence or when opaque, ‘black-box’ AI tools obscure the basis for recommendations and decisions. Even when the use of AI is disclosed and any final decision rendered by humans, the anchoring effect and undue deference to technology can change outcomes. Beyond all the practical issues is whether humans want important issues decided for them by machines.
This article concludes by looking at where advances in AI capabilities might take courts in the future. As technology advances and user resistance drops, AI will be available for more tasks. Using AI appropriately will require a nuanced, technically sophisticated approach. AI influence on the process will not be cured by ‘judge washing’ the algorithm with a final human decision maker. Just as ethical rules control and limit the influence other humans can have on judicial decisions, systems and rules will have to evolve for AI to make sure the influence of algorithms is appropriately controlled.
A realistic view of what role AI can play in the judicial process requires first a look at how AI operates and what functions it can be expected to perform in the near to intermediate term. In addition, in light of the developing field of judicial studies, we must look--as other commentators in this area have not--at the full scope of the judicial function, which goes far beyond just resolving individual cases. Only by matching the real potential of AI with the full range of judicial functions can we give a non-hyperbolic assessment.
Setting aside the possibility of radical technological advances, what we can expect in the near term is for software to play a role supporting--but not replacing--human judges. AI can, in certain cases, predict how a certain case might come out. That, however, falls far short of what judges do. The current capability of AI is limited to specialized tasks, and the roles of judges are so generalized that there is no near-term possibility of AI wholly and satisfactorily displacing judges in high stakes cases.
Even if software is developed to perform generalist capabilities, an uncertain prospect at best,4 we must still face the issue of whether we are prepared to delegate the creation and application of legal rights and responsibilities to impersonal artificial entities. In this regard it is important to remember that even if AI makes *325 the leap to general intelligence, such AI will in no way be human. Put simply, while in some ways as capable as human beings, AI will remain alien in fundamental ways. Whether such an intelligence can fulfill the diverse and fundamental roles played by human judges requires value choices beyond the scope of technology.
The uses for AI in chambers today include almost everything but the full delegation of significant cases. Overworked judges have admitted to using AI to be sure that their sense of normal outcomes is correct. Beyond ad hoc use, in China data analysis of the facts and law in a case allows judges and their supervisors to compare outcomes to predicted outcomes based on the dataset; this guides judges and also helps supervisors identify divergent results that might be due to incompetence or corruption. Many courts use AI to help guide the public as they encounter an unfamiliar legal system, whether that takes the form of robots greeting them at the courthouse entrance or guiding them step by step through online case submission. Transcript creation, document and transcript summaries, document review, AI-aided legal research, scheduling hearings, and translating documents are also among AI’s uses. More controversially, AI allows algorithmic recommendations as to whether pre-trial release or probation should be granted to criminals and what level of punishment should be imposed.
The appetite for AI adoption varies by jurisdiction. In his most recent annual message on the U.S. judiciary, Chief Justice Roberts praised the use of technology but also made clear that human decision-making was preferred. In China, by contrast, courts are required to implement a wide range of artificial intelligence tools by 2025. England and Wales has led in allowing non-lawyers, including AI-powered non-lawyers, to offer services to the public, while Singapore has embraced technology as a way to enhance its competitive appeal based on the rule of law.
This increased use of artificial intelligence in deciding cases raises both practical and philosophical questions. Biases are captured in data that reflect past, often discriminatory, practices, and these can infect AI recommendations. Due process concerns arise when any part of the right to a hearing is delegated to an alien, non-human intelligence or when opaque, ‘black-box’ AI tools obscure the basis for recommendations and decisions. Even when the use of AI is disclosed and any final decision rendered by humans, the anchoring effect and undue deference to technology can change outcomes. Beyond all the practical issues is whether humans want important issues decided for them by machines.
This article concludes by looking at where advances in AI capabilities might take courts in the future. As technology advances and user resistance drops, AI will be available for more tasks. Using AI appropriately will require a nuanced, technically sophisticated approach. AI influence on the process will not be cured by ‘judge washing’ the algorithm with a final human decision maker. Just as ethical rules control and limit the influence other humans can have on judicial decisions, systems and rules will have to evolve for AI to make sure the influence of algorithms is appropriately controlled.
A realistic view of what role AI can play in the judicial process requires first a look at how AI operates and what functions it can be expected to perform in the near to intermediate term. In addition, in light of the developing field of judicial studies, we must look--as other commentators in this area have not--at the full scope of the judicial function, which goes far beyond just resolving individual cases. Only by matching the real potential of AI with the full range of judicial functions can we give a non-hyperbolic assessment.
Setting aside the possibility of radical technological advances, what we can expect in the near term is for software to play a role supporting--but not replacing--human judges. AI can, in certain cases, predict how a certain case might come out. That, however, falls far short of what judges do. The current capability of AI is limited to specialized tasks, and the roles of judges are so generalized that there is no near-term possibility of AI wholly and satisfactorily displacing judges in high stakes cases.
Even if software is developed to perform generalist capabilities, an uncertain prospect at best,4 we must still face the issue of whether we are prepared to delegate the creation and application of legal rights and responsibilities to impersonal artificial entities. In this regard it is important to remember that even if AI makes *325 the leap to general intelligence, such AI will in no way be human. Put simply, while in some ways as capable as human beings, AI will remain alien in fundamental ways. Whether such an intelligence can fulfill the diverse and fundamental roles played by human judges requires value choices beyond the scope of technology.
In order to provide what I thought my students needed, I found it necessary to develop a textbook, which is available without charge at https://opencasebook.org/casebooks/7581-us-civil-procedure-for-international-students-2022-2023-edition/