See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/375558249

The possibilities, limitations, and dangers of generative AI in language

learning and literacy practices

Conference Paper · November 2023

CITATIONS READS

3 4,040

1 author:

Edwin Creely
Monash University (Australia)
78 PUBLICATIONS 699 CITATIONS

SEE PROFILE

All content following this page was uploaded by Edwin Creely on 11 November 2023.

The user has requested enhancement of the downloaded file.

 The possibilities, limitations, and dangers of generative AI in language learning and
literacy practices

Edwin Creely
Faculty of Education
Monash University
Melbourne Australia
edwin.creely@monash.edu

The rapid advancements in generative AI have produced significant interest in its potential
applications in language education. In this paper I examine the transformative possibilities of
AI-driven language learning tools, including amongst a range of possibilities, personalised
learning experiences, interactive content, and adaptive feedback, all of which can augment
language learning and literacy development. At the same time there needs to be acknowledge
of the limitations of generative AI in language education. Challenges like ensuring accuracy
and authenticity in AI-generated language content across diverse linguistic contexts are being
identified. Moreover, there are concerns about the potential for homogenising language usage
and endorsing narrow cultural representations through AI models, as well as issues related to
the dependency on technology and potential reduction in critical thinking and creativity.
Being aware of the dangers of generative AI in language learning, the paper also focuses on
the ethical implications, including issues to do with authentic authorship and control over
creative outputs. The significance of transparent and responsible AI use is stressed to address
concerns in higher education about the source of student language work. By scrutinising the
possibilities, limitations, and dangers of generative AI in language learning and literacy
practices, I aim to bring attention to the need to make informed decisions regarding AI
integration. The thrust of this paper is about ensuring a balanced approach that maximizes
the benefits of generative AI in language education while mitigating potential risks and
promoting responsible and inclusive language education in this emerging AI-driven digital
era.

Key words

generative AI, language education, literacy, ethics

Introduction

The domain of educational technology and the rise of artificial intelligence (AI) has
witnessed a series of transformative innovations over the last 30 years, tracing back to the
advent of the Internet in 1994 (Bozkurt et al., 2021; Creely, 2023; Luckin et al., 2016).
Arguably, the evolution of generative artificial intelligence (AI) is one of the more important
of these innovations, coming into public consciousness through ChatGPT (Trust et al., 2023;
Williams, 2023). AI systems based on sophisticated models like GPT-4 can produce text that
closely resembles human writing, making them essential assets for mastering languages.
Software powered by AI can offer intuitive feedback, devise practice activities, and construct
lifelike dialogue situations for language training. Studies indicate that AI holds promise in
enhancing language learning results (Liang et al., 2021). This technology embodies the
intersections of machine learning, training on language models, and deep learning based on
artificial neural networks. It has the capacity to generate original outputs such as documents
and graphical materials akin to human-like intelligence and creative production based on
natural language processing (NLP) and neural machine translation (Bown, 2012; Heaven,
2020; Lim, et al., 2023). As it continuously evolves, breaking barriers and setting new
standards, its influence on language education has taken centre stage, prompting enthusiasm,
critical discussions, and considerable scepticism (Liang et al., 2021).
Part of the enthusiasm for this new technology is the capacity of generative AI to
customise learning content and provide adaptive online and digital feedback may present a
significant leap for language education (Galaczi, 2023. Traditional language learning
methodologies, while effective in some instances, often follow a linear, one-size-fits-all
approach that may not be relevant for the digital age (Zhou & Niu, 2015). In contrast, AI-
driven tools can analyse individual learning patterns, adjust the curriculum accordingly, and
provide instant feedback in online and flipped learning environments (Su & Yang, 2023).
This can enhance the learning process but also ensures a more comprehensive understanding
of language, catering to individual learning styles and needs. Furthermore, these tools can
generate interactive content, enhancing sustained engagement. Imagine a virtual tutor capable
of crafting on-the-spot language activities and assessments, adjusting the teaching pace, or
even adapting the teaching strategy based on real-time student feedback. Such dynamic and
interactive experiences can make language acquisition more intuitive and enjoyable, driving
higher retention and fluency.
However, as with all technological innovations, the potential application of generative
AI in language education may have challenges. First, there is concern about the accuracy and
authenticity of the texts produced by generative AI. While an AI model might be accurate in
terms of grammar and syntax, the nuances of cultural meaning, including idioms inherent in
human languages, might be missing, or misrepresented (Pokrivcakova, 2019). Relying solely
on AI-generated content runs the risk of learning that embodies a standardised version of a
language, bereft of cultural richness and complexities. There might also be the potential
erosion of critical thinking and creativity. If learners become overly dependent on generative
AI bots, there is the potential for overly passive learning, and undue acceptance of AI-
generated responses without critical questioning. This could stifle analytical skills, creativity,
and even their ability to think in novel ways.
The ethical repercussions of using generative AI in language learning cannot be
overlooked. With the ability of AI to generate essays, articles, creative materials, translations,
and other forms of content, the lines between authentic student effort and AI-assisted outputs
become blurred. This suggests on-going questions about authenticity, originality, and the
human basis of learning (Craig, 2022). For example, should a student's AI-assisted writing
receive the same recognition as one written without the use of generative AI? More generally
there are ethical questions about the language sets on which generative AI is trained.
Potentially, generative AI offers significant possibilities for language education, but,
at the same time, there are also a set of challenges and ethical dilemmas. It is important to
find a balance in which there is the leveraging the learning potential of AI in language
education, along with caution, critical awareness, and respect for the essential humanity and
cultural basis of language.
In this position paper, the ideas identified above are explored towards a consideration
of how language educators, researchers and policy maker might conceive of the applications
of generative AI in language learning and classroom practices. A model to guide thinking in
this emerging area of research and practice in language education is offered as an outcome.
This writing is not designed to give practical applications as such but to engender thinking
and possibilities in this emergent space.

2
 Possibilities in AI-driven Language Learning

In this section, four affordances for the use of generative AI in language education are
discussed. These should be taken as expansive areas where generative AI might bring new
possibilities to language education in contexts such as second language learning and language
for academic purposes. Applications in this space are still being developed, so this discussion
should be considered quite speculative at this stage.

Affordance 1: Personalised Learning through Generative AI

Generative AI has emerged as a potential game-changer because of its ability to

deliver personalised learning experiences for language students. Historically, educational
pedagogies and approaches in language learning are often static, reflecting the standardisation
of curriculums (Zhou & Niu, 2015). These traditional approaches, while structured and
predictable, do not necessarily account for the unique learning styles, pace of learning, and
preferences of individual students. The result may be a system of language learning where
many learners were left behind or struggle to meet curriculum outcomes.
Potentially, with generative AI, this landscape of language learning might be
transformed. Unlike static educational approaches and tools, generative AI has the capability
to adapt in real-time, be part of the development of targeted learning materials, and be
integral to multimodal forms of content delivery and interactive learning (Baidoo-Anu &
Owusu Ansah, 2023; Vandewaetere & Clarebout, 2014). Generative AI can assist language
educators to design a learning path that is individualised as a sort of learning assistant.
Instead of constraining students to fit into a curriculum, the curriculum now fits the needs and
learning preferences of a student. This is built on the premise that students learn in different
ways and have preferences for how they would like to learn. For example, language learner
who prefers more visual content might be presented with a range of multimodal materials
such as infographics and charts. AI can help generate audio resources to construct bespoke
lessons for students who prefer audio-based instruction. Similarly, if a student excels in
vocabulary but struggles with grammar, the AI-driven bot can adjust its focus, dedicating
more resources and exercises towards improving grammatical skills through targeted
exercises.
Such a high degree of customisation ensures that students are not just passive
recipients of information but active and agential participants in their own learning journey.
This not only enhances their engagement levels but also promotes language retention and
understanding of their optimal ways of learning. Moreover, this personalised approach
potentially reduces feelings of inadequacy or frustration that may be present with more
traditional standardised approaches in language learning. When students experience real-time
progress tailored to their strengths and weaknesses, it is likely to enhance their confidence
and motivation.
Generative AI has foreshadowed a new expansive approach in language education,
moving away from broad-brush teaching approaches to intricate, individualised learning
experiences. It promises a future of customisation for every learner, ensuring a more effective
and enriching educational journey. However, changing well-established systems of teaching
and learning that are framed around a standardised curriculum is not likely to be easy. There
might even be resistance to the incursion of generative AI in the language education sector.
Unquestionably, it gives both the teacher and the student more agency in the learning process.

3
Affordance 2: Constructing Interactive Content in Language Learning

As the educational sector continues to change with technological advancements, one

of the standout benefits of generative AI is its ability to produce interactive content under the
direction of an educator (Bozkurt & Sharma, 2023; Tlili et al., 2023). This innovation may
reshape the learning landscape, particularly in language education, by transmuting passive
learning environments, with more linear types of learning exchanges, into dynamic, engaging
experiences.
Traditionally, language learning resources, be it textbooks or even digital content,
have been static. Students would consume the information and then, often separately, attempt
exercises or quizzes to test their understanding. Generative AI disrupts this paradigm by
integrating learning and assessment such that it becomes more fluid and personalised. For
instance, if a student is reading a passage and struggles with a particular word or phrase, the
AI can instantly detect this obstacle and provide elucidations or additional resources,
ensuring a seamless learning experience. Interactive content goes beyond merely presenting
information; it invites students to be a part of the learning process and invested in
collaborative learning (Kim et al., 2022). By responding to student inputs, generative AI
transforms lessons into two-way conversations that are more-or-less relational. Recent
research suggests that generative AI fosters a relational machine-human dynamic that goes
beyond merely seeing AI as a tool (Liu, 2021; Sanders & Wood, 2021).
For language learners, this might be quite impactful. Language is, at its core,
interactive and based in communicative outcomes. It's not just about memorising vocabulary
or grammar rules; it's about communication, dialogue, and expression. Generative AI's ability
to simulate such interactions makes learning more aligned with the real-world application of
language skills, such as conversation scenarios and culturally situated literacy practices
(Bozkurt, et al., 2023). With such interactive content, lessons become more explorative and
interactive. Students are encouraged to try out different sentence constructions, play with new
vocabulary, or even attempt to express complex ideas, knowing that the AI will guide and
correct them in real-time. This safe environment, where mistakes are immediately addressed,
encourages experimentation and risk-taking, which is crucial for language learning (Woo &
Choi, 2021). In addition, accounting for generative AI's capability for novelty, content can be
designed for expansive learning based on student responses. For instance, if a student excels
in a particular topic, the generative AI might proffer more advanced concepts or additional
challenges, such that the student is always pushed to learning possibilities beyond minimal
expectations.
The emergence of interactive content through generative AI also has implications for
educators and curriculum designers. Instead of creating linear lesson plans, educators can
now design modular, adaptive content. They can integrate various resources – videos,
quizzes, interactive exercises – that the AI can then select from based on student performance
and areas of strength and limitation.
Generative AI's capacity to produce interactive content in collaboration with an
educator marks a notable potential shift in the delivery and pedagogical basis of language
education. By changing passive consumption into active participation, it not only enriches
engagement but also has the potential to engender a holistic and effective learning experience
for students.

Affordance 3: Shifting Approaches to Feedback for Learning

Feedback has always been an integral part of the educational process. Whether it's a
teacher marking an assignment, a tutor providing verbal support, or an online platform

4
offering automated responses, feedback serves as the bridge between current understanding
and desired mastery. However, traditional feedback mechanisms, particularly in language
education, come with inherent limitations: they are often static, generalised, and delayed
(Inevich & Panadero, 2021). This is where generative AI, with its capability to provide
immediate, adaptive feedback, might profoundly alter the ways that educators offer feedback.
In a conventional classroom situation, a student might submit an assignment and wait
days, if not weeks, for it to be returned with comments. This delay potentially disrupts the
learning process, which, in language education, is especially important. The student has likely
moved on to other topics by the time feedback arrives, making it more difficult to return to
content and understanding that have passed. Furthermore, traditional feedback is often fixed.
It provides a snapshot of student performance but does not move and evolve as the student
progresses in their learning. This means that even as learners grow and change, the feedback
they receive remains rooted in their past performance, limiting its efficacy.
Generative AI migh facilitate a paradigm shift to this feedback process (Baidoo-Anu,
& Owusu Ansah, 2023; Su & Yang, 2023). Unlike its traditional counterparts, generative AI,
prompted and directed by a teacher, can analyse a student's work instantly and provide
feedback concurrent to the learning. For example, the simple task of writing a sentence and
then speaking it in a foreign language can be challenging for new learners, but generative AI
can provide instant feedback about grammatical errors and then practicing pronunciation and
being guided on-the-spot to correct misarticulations. This immediate feedback ensures that
mistakes are corrected as they happen, developing immediate learning and preventing the
solidification of errors over time.
Indeed, the feedback provided by generative AI is adaptive and potentially self-
learning as it is presented with additional language data. As the AI bot continuously learns
from a student's interactions, it can modify its feedback in response to the state of a learner's
progression at a point in time or for content areas in a curriculum. For example, if a student
constantly struggles with a particular verb tense in a language, the AI might generate more
detailed explanations and language exercises related to that concept.
Though at the present time there is little research to support this assertion, with
immediate and adaptive feedback using generative AI, the language learning process might
be significantly expedited. For instance, students are no longer required to wade through
detailed content to address their learning issues or wait for feedback to gauge their
understanding. Instead, they are in a constant loop of learning, practicing, receiving feedback,
and refining – a cycle that enhances retention and mastery. Such a feedback system also
promotes learner autonomy. Students can take charge of their learning journey, adjusting
their pace and approach based on the continuous insights they receive.
By replacing static, delayed feedback with a dynamic, real-time system, it addresses
one of the long-standing challenges in education: the immediacy of feedback at the point of
learning. For language learners, overcoming this challenge holds the promise of faster, more
efficient, and more personalised learning experiences, based on feedback that is not just
retrospective but also immediate and forward-looking. This has implications for language
educators and how they use and work with generative AI to support the immediacy of this
learning.

Affordance 4: Augmenting Human Creativity in Text and Graphics

Generative AI stands at the intersection of technology and creativity, hypothetically

fostering a new era where human imagination and creativity can be amplified, disrupted and
extended by machine intelligence (De Cremer et al., 2023; Moruzzi, 2021). This synergy is
particularly evident in textual and graphical content creation. In conventional approaches to

5
writing processes, authors draw upon their knowledge, experiences, and intuition to craft a
range of textual genres, including imaginative and personal forms of writing with the aid of
generative AI, writers can explore a wider array of ideas, styles, and perspectives. The AI can
suggest alternate phrasings, offer novel language possibilities, expand on initial ideas, or even
generate entirely new content based on the parameters set through prompt engineering (?).
Such collaboration between generative AI and a learner can facilitate a richer, more diverse
creative process, where human creativity is complemented by the vast language-based and
data-driven capacities of AI.
Beyond textual content, generative AI has the capacity to create graphical content in
response to both textual and visual input (Nah et al., 2023; Roose, 2022). Educational
designers can feed initial sketches or concepts into an AI system, which then produces a
range of design variations, patterns, or colour schemes. This collaborative approach
accelerates the design process and opens avenues for novel visuals materials in that can then
be used human designers in language learning contexts.
In essence, generative AI acts as a creative partner in language education, offering
fresh perspectives and ideas, alongside new textual and graphical resources that can be
utilised by educators and educational designers (Schober, 2022). By blending human creative
ingenuity with AI's computational power, there is expansive possibilities and augmentation
that go beyond the boundaries of traditional ways of producing original learning content and
student textual and graphical outputs (Horvatić & Lipic, 2021; Peeters et al., 2021).

Challenges of using generative AI

While the possibilities with generative AI are significant, it is crucial to recognise and
address the inherent challenges of generative AI in language education, as part of a balanced
and critical approach to the possibilities for implementation. Generative AI is being
positioned as an indispensable tool in contemporary education internationally, potentially
challenging traditional pedagogies and offering personalised, interactive experiences for
learners (Humble & Mozelius, 2019; Tzirides et al., 2023).

Challenge 1: Considering the Issue of Accuracy and Authenticity

A pressing concern in AI-driven language content is ensuring both accuracy and

authenticity, especially given the linguistic complexities and pragmatics that characterise
human languages (Stokel-Walker & Van Noorden, 2023). Accuracy, in the context of
language, generally refers to the grammatical and syntactical correctness of content.
Generative AI has shown initial impressive results across a range of work and education
sectors, based on the vast datasets and complex algorithms that drive its potential
(Brynjolfsson et al., 2023). AI systems can generate content that often adheres to the rules of
grammar with exactness. For students and educators in language education, this precision
offers a valuable tool for understanding the structural aspects of a language and for precise
translation. However, language is not merely a set of linguistic rules but has situatedness and
context in which it is embedded. It reflects culture, history, and shared experiences which are
subject to change. This is where the challenge of authenticity comes into play. Authenticity
encompasses the nuances, idioms, and cultural contexts that give a language its unique set of
flavours and embodies the difference between understanding words in a sentence and
grasping the sentiment and contextual purpose behind them (MacDonald et al., 2006).
For instance, while a generative AI system might well flawlessly translate an idiom or
form of expression literally from one language to another, it might not represent the cultural,
historical, or social context behind that idiom. Such gaps of meaning can lead to a surface-

6
level understanding of a language that lack the deeper connections that are pivotal to the
functional learning of language for rich communication. The risk, then, is twofold. First,
learners might end up with a sterile, homogenised, and sanitised version of a language,
missing out on its richness and its cultural specificities, which has been a potential issue with
English language learning (Kushner, 2003). Second, there's the potential for
miscommunication. Without understanding cultural nuances, learners might use words or
phrases appropriately in a grammatical sense, but inappropriately in a cultural context.
Addressing these challenges requires a multi-faceted approach. While generative AI can
serve as a foundation, human oversight and direction are decisive to maintain authenticity.
Collaborative systems, where AI-generated content is reviewed and enriched by human
linguists or educators, might be a way forward.
As generative AI continues to play a larger role in language education, striking a
balance between accuracy and authenticity will be imperative. Only by acknowledging and
addressing this challenge can educator realistically harness the potential of AI in the
development of holistic language learning.

Challenge 2: The Issue of Normalised Language from AI Models

The transformative role of AI in various sectors, including in language education, has

accelerated over the last two years with the strong emergence of generative AI in its various
iterations (Hao & Yunyun, 2023; Köstler & Ossewaarde, 2021). The efficacy of generative
AI across languages other than English is still being evaluated (Ahuja et al., 2023). As with
all technological advancements, there are attendant concerns. As noted above, one significant
concern is the potential homogenisation of language usage through AI models. As AI systems
get trained on vast, often mainstream and western datasets, there's a risk that the language
they generate and teach might become overly standardised, sidelining the rich diversity that
characterises human languages, sublanguages and dialects. Language, in its essence, is not
just a communication process supported by linguistic structures but a reflection of culture,
heritage, and identity. Each region, community, and even individual brings a unique flavour
to language through dialects, colloquialisms, and cultural idioms (Yağiz & Izadpanah, 2013).
Languages capture histories, traditions, emotions, and shared experiences which are beyond
the scope of normalisation.
When educators and learners overly rely on AI models, they might be introduced to a
sanitised or authorised version of the language, one that adheres to grammatical rules but
might be stripped of regional variations and colloquial richness. For instance, the English
spoken in London is distinct from that in New York, Mumbai, Hanoi, or Melbourne. Each
version of English carries its set of idiomatic expressions, slang, and pronunciation
differences, rooted in the region's culture and history. Care is needed so that learners engage
with the cultural context of the language with its rich understandings. As language tools
based on generative AI become more prevalent, there's a risk of these regional variations will
diminish, leading to a loss of linguistic diversity. To counter this, it may well be essential for
AI developers and linguists to work collaboratively. AI models should be trained on diverse
national and cultural datasets to capture the myriad variations of a language. However, the
language models for the training of generative AI tend to be western at the current time.
Additionally, language learning platforms might incorporate modules focusing specifically on
regional dialects and colloquialisms, ensuring learners get a holistic understanding of
linguistic diversity.
While AI might offer unparalleled advantages in language learning, it is important to
remain vigilant against the inadvertent homogenisation of language. Preserving linguistic

7
diversity is essential for keeping languages vibrant, relevant, and deeply connected to their
cultural roots.

Challenge 3: Maintaining Critical Thinking and Creativity

Another pressing challenge is the potential erosion of critical thinking and creativity
among students due to over-reliance on AI technologies (Bown, 2012; de Vasconcellos et
al., 2021). At the heart of education is the goal to promote an inquisitive spirit, encouraging
students to question, challenge, and create. Critical thinking equips learners to dissect
information, understand underlying concepts, and form independent conclusions. Creativity,
on the other hand, propels them to imagine, innovate, and express original thoughts. These
sets of skills and understandings are essential not just for academic success but for navigating
the complexities of the contemporary world with its digital culture (Padget, 2013).
When students become overly dependent on AI tools, there is a risk that they may
shift into a more passive learning mode, which is the other side of the coin to the affordances
around agency discussed above. For instance, if generative AI consistently provides answers
to complex problems, corrects errors or auto-generates writing, students might sidestep the
rigorous cognitive process that would typically be required. They might start viewing the
output of AI as the definitive answer, curtailing their own analytical thinking and inhibiting
the exploration of alternative solutions or perspectives. In all this turn to generative AI the
power of human criticality cannot be undermined.
Furthermore, research suggests that creativity flourishes in learning environments
where norms and boundaries are challenged and there is risk-taking but in safe ways (Creely
et al., 2021). If AI-generated content becomes the accepted norm, students may shy away
from thinking outside the box, believing that he way of AI is the only way. This could stifle
original thought and discourage students from challenging boundaries in terms of creativity.
To navigate these challenges, a balanced approach is imperative. Educators should
emphasise the supplemental role of AI, positioning it as a technology to enhance, not replace,
human capacities. Engaging language students in debates, brainstorming sessions, and open-
ended projects can promote critical thinking and creativity, ensuring that these crucial
capacities remain at the forefront of education and include generative AI as an expansive
possibility rather than diminishing human agency. While AI holds immense potential in
reshaping education, it is crucial to use it judiciously. Fostering an environment where critical
thinking and creativity are celebrated will ensure that students are prepared not just for
academic pursuits but for the challenges and opportunities of the broader world.

Ethical Implications

The integration of generative AI in language education is not without its ethical implications.
In this section three possible ethical implications are discussed.

Implication 1: Authentic Authorship and generative AI

The rise of generative AI and its seemingly endless potential for content creation
presents both opportunities and dilemmas. A paramount concern is the blurred distinction or
over-lapping between human-generated and AI-produced content, particularly in academic
and professional settings (Marr, 2023). As generative AI systems gain increasing proficiency
and scope in producing a range of textual and graphical materials, the question of authentic
authorship becomes increasingly pertinent.

8
 When an educator, student or writer uses AI to generate or refine content, it
challenges traditional notions of originality. The question is this: Who ‘owns’ the work? Is it
the individual who provided the initial input through a prompt, or the AI that shaped and
fleshed out the content? (Craig, 2022). This not only has ethical implications but also impacts
the value and integrity of educational assessments and published materials (Baidoo-Anu &
Owusu Ansah, 2023; Qadir, 2023). If a student offer work for assessment, how is it to be
assessed if AI figured in its production? This might mean a reconsideration of how educators
in the language space conceive of assessment and what is assessed, even the modes of
assessment.
As educators, designers and students integrate AI into content creation, it's imperative
to establish clear guidelines and ethical standards. Recognising and crediting collaborative
efforts with AI tools ensures transparency and upholds the inviolability of authentic
authorship in this rapidly evolving digital landscape. Then again, is authorship with
generative AI also authentic?

Implication 2: Human Creativity and AI in Content Creation

The integration of AI in content creation sparks a compelling debate about ownership

and originality in educational contexts (Su & Yang, 2023). When a student employs AI for
content assistance, delineating the boundary between human ingenuity and machine output
becomes complex. The essence of creativity typically rests with the creator, but with AI's
involvement, the lines become clouded. Is the final piece a product of the student's vision, the
AI algorithm, or a fusion of both? This convergence raises crucial questions about the nature
of creative ownership, suggesting the need to re-evaluate and redefine authorship in the AI-
augmented age.

Implication 3: Integrity through Transparent AI Use in Education

In higher education and across all sectors in education, including in language

education, the infusion of AI demands a commitment to transparency (Zhuo et al., 2023). As
generative AI become more integrated into academic processes, there is a clear mandate for
institutions to communicate their usage. This ensures that both educators and students
understand when AI interposes and to what extent. However, it must also be stated that such a
mandate in a rapidly moving technological space may not be easy.
Additionally, for assessments and assignments, a transparent system can bring some
level of veracity and authenticity regarding student work, distinguishing between human
creativity and AI-generated content (Crawford et al., 2023). By championing transparency
and responsible AI use, education institutions can maintain academic integrity while
embracing technological advancements (Mhlanga, 2023a, 2023b). Critically, however,
finding lines of demarcation between machine and human generated content may not be
simple, and it could be the case that it is not possible, depending on the type of output.

Conceiving of generative AI in language learning

This section presents a conceptual or thinking model based on the concepts introduced
in this paper. The model may be useful for researchers, practitioners and policy makers in the
language education space when considering how generative AI might be adopted and used
appropriately and ethically. This model is presented in Figure 1.
Figure 1 visualises the intersection between Generative AI, its functionalities in
language, and its implications for language education. Starting with "Generative AI," this

9
section emphasises the technology's capabilities in “Language functionality”. AI, especially
when generative in nature, can accomplish a wide range of tasks related to language and to
novel outputs from training on language models. These include translating between
languages, correcting errors, summarizing content, expanding on ideas, creating new content,
responding to queries, and sourcing information, to identify but a few. These functionalities
attest to the potential of AI to assist users in numerous language-related tasks.
Moving to the central portion, "Areas for research, policy, and practice," questions are
posed that call for a deeper exploration of AI's role. "What are the affordances?" seeks to
understand the benefits and capabilities AI brings. "What issues need to be considered?"
prompts stakeholders to contemplate ethical, practical, and technological concerns. Finally,
"What are possible futures?" encourages forward-thinking, envisioning how AI might evolve
and further integrate into our daily lives and educational systems.

Figure 1. A model for generative AI in language learning

The last section, "Language education," suggests the end application of these
considerations of transformation. It signifies that the integration of Generative AI's
functionalities can revolutionize how we approach language learning, making it more
dynamic, personalised, and efficient. The entire diagram underscores the importance of
understanding AI's capabilities, especially its capacity for novelty, independent content
creation, and potential for adaptability, feedback, and interactivity. The diagram points to the
need for thoughtful and critical integration into the core practices in language education. It
also suggests the need for AI literacies by educators and their students.

Conclusion

The digital transformation of education, intensified by the rise of generative AI,

signifies a decisive moment in the future of language learning. The promises held by this
technology are vast: personalised content, instantaneous feedback, immersive, interactive
learning experiences, and content creation. This technology might revolutionise how
languages are taught and learned, potentially making the process more efficient, engaging,
and customised as a response to individual needs.
At the same time, the ability of generative AI to emulate human-like thought
processes and generate content, also presents profound challenges. The essence of language is

10
deeply rooted in culture, emotion, and human experience. While AI can master syntax and
grammar, capturing the human and contextual dimension of language is more problematic.
The risk of delivering a homogenised version of a language, stripped of its cultural
intricacies, is a possibility that all stakeholders in language education should consider.
Additionally, the question of agential and sustained learning emerges. The ease with
which AI can produce content could inadvertently reduce the effort students invest,
potentially diminishing critical thinking and human creativity. Relying heavily on AI might
provide immediate results, but it also could deprive learners of the challenges which lead to
deep learning.
The ethical considerations surrounding AI-assisted outputs further complicates the
integration of the technology. With the possibility of hybridity between student effort and AI
generation, educational institutions face challenges in assessment and recognising genuine
work, and perhaps new ways of thinking about the problem need to be developed. Generative
AI proffers a paradigm shift in how educators and policy makers define, measure, and value
academic achievement. In addition, the advent of AI points to the need for robust measures to
ensure data privacy, transparency, and ethical usage. Embracing AI's potential should not
mean forsaking the core values of education. A harmonised approach, where AI complements
human effort rather than replacing it, seems prudent. Institutions must also prioritise
transparency, ensuring all stakeholders understand the extent and implications of AI's role.
Finally, the integration of generative AI into language education offers a new learning
horizon full of opportunities, but it is also a horizon that must be envisioned with care,
respect, and foresight. Balancing the revolutionary capabilities of AI with respect for the
values of genuine learning and cultural understanding will determine the success of this
endeavour, and this requires critical literacies and substantial understanding of its potential
and pitfalls. The thinking or conceptual model offered in the previous section is one attempt
to bring the sort of understanding in this space that can lead to positive and sustainable
change.

References

Ahuja, K., Diddee, H., Hada, R., Ochieng, M., Ramesh, K., Jain, P., Nambi, A., Ganu, T.,
Segal, S., Axmed, M., Bali, K., & Sitaram, S. (2023). MEGA: Multilingual evaluation of
generative AI. arXiv:2303.12528. https://doi.org/10.48550/arXiv.2303.12528
Baidoo-Anu, D., & Owusu Ansah, L. (2023). Education in the era of generative artificial
intelligence (AI): Understanding the potential benefits of ChatGPT in Promoting
teaching and learning. https://dx.doi.org/10.2139/ssrn.4337484
Boden, M. (2018). Artificial intelligence. A very short introduction. Oxford University Press.
Bown, O. (2012). Generative and adaptive creativity: A unified approach to creativity in
nature, humans and machines. In J. McCormack, J. & M. d’Inverno, M. (Eds),
Computers and Creativity (pp. 361–381). Springer.
Bozkurt, A., Karadeniz, A., Baneres, D., Guerrero-Roldán, A., & Rodríguez, M. (2021).
Artificial intelligence and reflections from educational landscape: A review of AI studies
in half a century. Sustainability. 13(2). https://doi.org/10.3390/su13020800
Bozkurt, A. (2023). Generative artificial intelligence (AI) powered conversational
educational agents: The inevitable paradigm shift. Asian Journal of Distance Education,
18(1). https://doi.org/10.5281/zenodo.7716416
Bozkurt, A., & Sharma, R. C. (2023). Challenging the status quo and exploring the new
boundaries in the age of algorithms: Reimagining the role of generative AI in distance

11
 education and online learning. Asian Journal of Distance Education, 18(1), i-viii.
https://doi.org/10.5281/zenodo.7755273
Brynjolfsson, E., Li, D., & Raymond, L. R. (2023). Generative AI at work (Working Paper
No. 31161). National Bureau of Economic Research. http://www.nber.org/papers/w31161
Craig, C. (2022). The AI-Copyright challenge: Tech-neutrality, authorship, and the public
interest. All Papers. https://digitalcommons.osgoode.yorku.ca/all_papers/360
Crawford, J., Cowling, M., & Allen, K. (2023). Leadership is needed for ethical ChatGPT:
Character, assessment, and learning using artificial intelligence (AI). Journal of University
Teaching & Learning Practice, 20(3). https://doi.org/10.53761/1.20.3.02
Creely, E., Henriksen, D., Crawford, R. & Henderson, M. (2021). Exploring creative risk-
taking and productive failure in classroom practice. A case study of the perceived self-
efficacy and agency of teachers at one school. Thinking Skills and Creativity,
42. https://doi.org/10.1016/j.tsc.2021.100951.
Creely, E. (2023). Conceiving creativity and learning in a world of artificial intelligence: A
thinking model. In D. Henriksen & P. Mishra (Edits.), Creative provocations:
Speculations on the future of creativity, technology & learning (pp. 35-50).
Springer. https://doi.org/10.1007/978-3-031-14549-0
De Cremer, D., Morini Bianzino, N., & Falk, B. (2023, April 13). How generative AI could
disrupt creative work. Harvard Business Review. https://hbr.org/2023/04/how-
generative-ai-could-disrupt-creative-work
de Vasconcellos, S., da Silva Freitas, J., Junges, F. (2021). Digital capabilities: Bridging
the gap between creativity and performance. In S. Park, M., Gonzalez-Perez, & D.
Floriani (Eds), The Palgrave Handbook of Corporate Sustainability in the Digital
Era. Palgrave Macmillan, Cham. https://doi.org/10.1007/978-3-030-42412-1_21
Galaczi, E. (2023). English language education in the era of generative AI: our perspective.
Cambridge University Press and Assessment.
https://www.cambridgeenglish.org/Images/685411-english-language-education-in-the-
era-of-generative-ai-our-perspective.pdf
Hao,Y., & Yunyun, G. (2023). Generative artificial intelligence empowers educational
reform: current status, issues, and prospects. Frontiers in Education, 8
https://www.frontiersin.org/articles/10.3389/feduc.2023.118316
Heaven, W. (2020, July 20). OpenAI’s new language generator GPT-3 is shockingly good—
and completely mindless. MIT Technology Review.
https://www.technologyreview.com/2020/07/20/1005454/openai-machine-learning-
language-generator-gpt-3-nlp/
Horvatić, D., & Lipic, T. (2021). Human-Centric AI: The Symbiosis of Human and Artificial
Intelligence. Entropy, 23(3), 332. MDPI AG. http://dx.doi.org/10.3390/e23030332
Humble, N., & Mozelius, P. (2019). Artificial intelligence in education - a promise, a threat
or a hype? European Conference on the Impact of Artificial Intelligence and Robotics,
EM-Normandie Business School Oxford, UK, 31 October-1 November 2019,
ECIAIR.19.005 (pp.149-156). https://doi.org/10.34190
Inevich, A. A., & Panadero, E. (2021). A review of feedback models and theories:
Descriptions, definitions, and conclusions. Frontiers in Education, 6.
https://doi.org/10.3389/feduc.2021.720195
Kim, J., Lee, H. & Cho, Y.H. (2022). Learning design to support student-AI collaboration:
perspectives of leading teachers for AI in education. Education and Information
Technologies, 27, 6069–6104. https://doi.org/10.1007/s10639-021-10831-6
Köstler, L., & Ossewaarde, R. (2021). The making of AI society: AI futures frames in
German political and media discourses. AI & Society (2021).
https://doi.org/10.1007/s00146-021-01161-9

12
Kushner, E. (2003). English as global language: Problems, dangers,
opportunities. Diogenes, 50(2), 17-23. https://doi.org/10.1177/0392192103050002002
Liang, J.-C., Hwang, G.-J., Chen, M.-R. A., & Darmawansah, D. (2021). Roles and research
foci of artificial intelligence in language education: An integrated bibliographic analysis
and systematic review approach. Interactive Learning Environments.
https://doi.org/10.1080/10494820.2021.1958348
Lim, W. M., Gunasekara, A., Pallant, J. L., Pallant, J. I., & Pechenkina, E. (2023). Generative
AI and the future of education: Ragnarök or reformation? A paradoxical perspective from
management educators. The International Journal of Management Education, 21(2),
100790. https://doi.org/10.1016/j.ijme.2023.100790
Liu, B. (2021). In AI we trust? Effects of agency locus and transparency on uncertainty
reduction in human–AI Interaction. Journal of Computer-Mediated Communication, 26
(6), 384–402. https://doi.org/10.1093/jcmc/zmab013
Marr, B. (2023, May 16). The Danger of AI Content Farms. Forbes.
https://www.forbes.com/sites/bernardmarr/2023/05/16/the-danger-of-ai-content-
farms/?sh=622f8caf4fca
MacDonald, M., Badger, R., & Dasli, M. (2006). Authenticity, culture and language learning.
Language and Intercultural Communication, 6(3-4), 250-261.
http://wrap.warwick.ac.uk/78443
Mikalef, P. & Gupta, M. (2021). Artificial intelligence capability: Conceptualization,
measurement calibration, and empirical study on its impact on organizational creativity
and firm performance. Information & Management, 58(3).
https://doi.org/10.1016/j.im.2021.103434.
Mhlanga, D. (2023a). Open AI in education, the responsible and ethical use of ChatGPT
towards lifelong learning. In FinTech and Artificial Intelligence for Sustainable
Development Palgrave Macmillan. https://doi.org/10.1007/978-3-031-37776-1_17
Mhlanga, D. (2023b). Open AI in Education, the Responsible and Ethical Use of ChatGPT
Towards Lifelong Learning. In FinTech and Artificial Intelligence for Sustainable
Development. Sustainable Development Goals Series. Palgrave Macmillan.
https://doi.org/10.1007/978-3-031-37776-1_17
Moruzzi, C. (2021). Measuring creativity: an account of natural and artificial
creativity. European Journal for Philosophy of Science, 11(1).
https://doi.org/10.1007/s13194-020-00313-w
Nah, F. F.-H., Zheng, R., Cai, J., Siau, K., & Chen, L. (2023). Generative AI and ChatGPT:
Applications, challenges, and AI-human collaboration. Journal of Information
Technology Case and Application Research, 25(3), 277-304.
https://doi.org/10.1080/15228053.2023.2233814
Padget, S. (Ed.). (2013). Creativity and critical thinking. Routledge.
Peeters, M., van Diggelen, J., van den Bosch, K., Bronkhorst, A., Neerincx, M., Schraagen, J.
& Raaijmakers, S. (2021). Hybrid collective intelligence in a human–AI society. AI &
Society, 36, 217–238. https://doi.org/10.1007/s00146-020-01005-y
Pokrivcakova,S. (2019). Preparing teachers for the application of AI-powered technologies in
foreign language education. Journal of Language and Cultural Education,7(3) 135-
153. https://doi.org/10.2478/jolace-2019-0025
Qadir, J. (2023). Engineering education in the era of ChatGPT: Promise and pitfalls of
generative AI for education. In 2023 IEEE Global Engineering Education Conference
(EDUCON) (pp. 1-9). Kuwait, Kuwait.
https://doi.org/10.1109/EDUCON54358.2023.10125121

13
 Roose, K. (2022, October 21). A.I.-generated art is already transforming creative work. The
New York Times. https://www.nytimes.com/2022/10/21/technology/ai-generated-art-jobs-
dall-e-2.html
Sanders, N., & Wood, J. (2021). Combining humans and machines in an emerging form of
enterprise: the Humachine. Foresight: The International Journal of Applied Forecasting,
61, 28–35.
Schober, R. (2022). Passing the Turing test? AI generated poetry and posthuman creativity.
Artificial intelligence and human enhancement, 21, 151.
https://doi.org/10.1515/9783110770216-009
Stokel-Walker, C., & Van Noorden, R. (2023, February 9). The promise and peril of
generative AI. Nature, 614.
https://fens.sabanciuniv.edu/sites/fens.sabanciuniv.edu/files/2023-07/the-promise-and-
peril-of-generative-ai.pdf
Su, J., & Yang, W. (2023). Unlocking the power of ChatGPT: A framework for applying
generative AI in education. ECNU Review of Education, 6(3), 355-366.
https://doi.org/10.1177/20965311231168423
Tlili, A., Shehata, B., Adarkwah, M. A., Bozkurt, A., Hickey, D. T., Huang, R., &
Agyemang, B. (2023). What if the devil is my guardian angel: ChatGPT as a case study
of using chatbots in education. Smart Learning Environments, 10(1), 1-24.
https://doi.org/10.1186/s40561-023-00237-x
Trust, T., Whalen, J., & Mouza, C. (2023). Editorial: ChatGPT: Challenges, opportunities,
and implications for teacher education. Contemporary Issues in Technology and Teacher
Education, 23(1). https://citejournal.org/volume-23/issue-1-23/editorial/editorial-chatgpt-
challenges-opportunities-and-implications-for-teacher-education
Tzirides, A.O., Saini, A., Zapata, G.C., Searsmith, D., Cope, B., Kalantzis, M., Castro, V.,
Kourkoulou, T., Jones, J., Silva, R.A., Whiting, J.K., & Kastania, N.P. (2023). Generative
AI: Implications and applications for education. ArXiv, abs/2305.07605.
https://doi.org/10.48550/arXiv.2305.07605
Vandewaetere, M., & Clarebout, G. (2014). Advanced technologies for personalized learning,
instruction, and performance. In J. Spector, M. Merrill, J. Elen, & M. Bishop (Eds.),
Handbook of Research on Educational Communications and Technology. Springer.
https://doi.org/10.1007/978-1-4614-3185-5_34
Williams, T. (2023, March 23). GPT-4’s launch ‘another step change’ for AI and higher
education. The Times Higher Education.
https://www.timeshighereducation.com/news/gpt-4s-launch-another-step-change-ai-and-
higher-education
Woo, J. H., & Choi, H. (2021). Systematic review for AI-based language learning tools.
arXiv:2111.04455 [cs.CY]. https://doi.org/10.48550/arXiv.2111.04455
Yağiz, O., & Izadpanah, S. (2013). Language, culture, idioms, and their relationship with the
foreign language. Journal of Language Teaching and Research, 4(5), 953-957.
https://doi.org/10.4304/jltr.4.5.953-957
Zhou, G., & Niu, X. (2015). Approaches to language teaching and learning. Journal of
Language Teaching and Research, 6(4), 798-802.
https://dx.doi.org/10.17507/jltr.0604.11
Zhuo, T. Y., Huang, Y., Chen, C., & Xing, Z. (2023). Exploring AI ethics of ChatGPT: A
diagnostic analysis. arXiv. https://doi.org/10.48550/arXiv.2301.12867

14

View publication stats