The transformative impact of artificial intelligence on higher education: A critical reflection on current trends and futures directions

Abstract

Artificial intelligence (AI) seems to offer considerable potential in higher education by offering personalized learning experiences and fostering deeper student engagement. However, whether the AI’s impact reflects historical patterns of technology integration in education, which may be modest and determined by the multiple education or research “functions”, or whether the techno-scientific breakthrough signifies a deeper transformation in higher education. This article explores the complexities of AI in education, separating promotional narratives from reality. It examines the role of AI in reshaping knowledge and learning, while warning against the loss of critical approaches. AI’s potential to transform education remains largely unrealized. Current applications reinforce existing inequalities and fail to deliver complete personalization. However, the picture is complicated, as AI’s impact on creative teaching, foreign language learning, and research-intensive disciplines is notable. AI represents a significant step towards blurring the boundaries between formal and informal education, and the wide “distribution” of both learning and knowledge (in symbiosis with other digital technologies). AI’s influence is growing, but its full potential in education has yet to be realized, and there’s no guarantee that it will be realized to the extent of techno prophetic hopes.

When talking about the impact of AI on education and higher education in particular, the assumptions are mainly focus on the promises, rather than the threats or ambiguities of it’s dramatic development, even if the difficulties are not ignored at all. Many authors or reports start by considering that AI is already bringing about major changes and will do so increasingly (Kumar et al., 2024). Collaboration between AI and humans in learning and research is seen as likely to bring substantial benefits to individuals and society. However, these benefits depend on ensuring that critical thinking skills and academic integrity remain at the forefront of the transformation that most analysts are seeing or calling for. Whether AI is generative, narrowly focused as it is today, or becomes multi-purpose, education should, by all accounts, foster creativity and improve problem-solving skills (Al-Zahrani & Alasmari, 2024).

The multiple benefits of introducing AI tools into Higher Education are tempered by the inherent risks of bias (such as skewed data) and potential fraud in both learning assessment and research processes and data (plagiarism or misuse of AI-generated content). ¹ AI tools can be used in problematic ways, raising wider ethical, political and economic questions. From this point of view, the dominant role of large technology companies in the development of AI further complicates the landscape, as these companies have vested interests in the adoption and expansion of the technology, which do not at all automatically converge with the common good, and which would require careful and critical scrutiny, which to date is only marginally engaged.

For all that, there is consensus on the crucial role that educators - teachers, researchers and university administrators - must play in ensuring the “responsible” integration of AI into education. This presupposes, first and foremost, professional development and ongoing training for higher education professionals - differentiated according to function - as well as active collaboration between them and AI experts from laboratories and industry. By working together, educators and industry engineers and researchers can help realize the potential of AI to improve learning outcomes and advance knowledge production.

In higher education precisely, AI is seen as having tremendous potential: it can be harnessed not only to generate content and personalize learning experiences, but also to foster deeper student engagement (Keep & Brown, 2018; Kumar et al., 2024). For example, AI-driven learning platforms can tailor exercises to students’ individual needs, providing personalized feedback that encourages active learning. At a more meta-analytical level, AI can also encourage students to critically interpret the information presented to them, deepening their understanding and promoting intellectual rigor. For example, AI tools such as plagiarism checkers or data analysis algorithms can help students reflect on the ethical dimensions of their research or check the reliability of their sources.

One question remains: do these quite optimistic visions of the role of AI in education, and primarily in higher education, correspond to historical patterns of technology integration? Examining the history of technological advances in education, as well as massification and current intensification of international competition in higher education, raises some other important questions. Are we witnessing a profound institutional transformation, or fragmented practices that are only partially integrated into the learning and teaching process? This article aims to explore these crucial questions, by assessing whether the voluntarist discourses most often valorizing the highly promising impact of AI in learning within higher education and research reflect and anticipate the complex realities of its implementation.

To understand and answer those questions, we will try to separate myths and wishful thinking or promotion discourses from what we really and accurately know, coming from social Sciences understanding of technology and their complex links with society, organizations, education or Knowledge production and circulation. So first, we will examine the role of science and technology such as AI, digital or digital technologies in helping to reshape, knowledge, education and culture. Considering these changes not only as an evolution in human learning devices but as a possible signal for a new phase in the exercise of rationality and the developments of massively shared irrationalities. In education and research, the intentional or de facto integration of technological tools, including today’s AI, has standardized the production and processes of knowledge production or learning/teaching, while raising concerns about the loss of critical approaches. However, emerging interdisciplinary efforts, including the possibility of defining ‘techno-digital humanism’, seek to balance technological advances with socio-philosophical and artistic reflections, potentially fostering a deeper, more holistic understanding of knowledge in a complex, ever-changing world.

Then, we’ll recall that Higher education faces huge pressure to adopt and promote AI in research, training, and administration. As many reports and commentaries point out, AI is expected to personalize learning, automate administrative tasks, and support research (OECD (2023), improving student outcomes and institutional efficiency. Meanwhile, Universities are urged to collaborate with industry to stay updated on AI developments and equip students with relevant skills, continuing a longstanding trend of aligning education with technological and economic needs.

Third, we show that Artificial intelligence, despite influencing higher education by shaping teaching, research, and management, is not the sole driver of transformation and is used strategically to ease professional pressures and enhance competitiveness for innovation, which are the main higher Education for decades. One must keep in mind, that Higher Education is undergoing massification, with prestigious universities setting global standards and others focusing on local applications. Technoscience, digital tools and AI reshape learning or research, under the umbrella of management models which emphasize innovation and economic growth.

The promise of a radical transformation of education through AI is still far from reality. Education systems are complex, and technology alone cannot address the diverse goals and constraints of human learning, (Sawyer, 2022), nor research, service to society or teaching activities. Educational promises of AI, such as fully personalized learning, are still to come. Plus, technologies such as AI and MOOCs often reinforce existing inequalities, benefiting those already equipped with the skills needed for autonomous learning and leaving others behind. What’s more, in fields like language learning, where AI tools have been in use for decades, progress remains limited. Most educational programs follow linear, one-size-fits-all approaches with minimal personalization.

But, even if the promise of a revolution is far from reality, (generative) AI is already part of a wider digital ecosystem, increasingly transforming education. AI tools, with their multimedia capabilities and user-friendly interfaces, can indeed benefit beginner learners in various disciplines such as language learning, mathematics and computer science, etc. These tools also allow students to practice anonymously, minimizing judgment and embarrassment.

For sure, AI blurs the boundaries between formal education and informal learning, offering more accessible educational experiences for individuals. But, higher education institutions, particularly in areas such as assessment or graduation, are slow to embrace these changes. In research-intensive disciplines, such as Masters and PhD programs, the impact of AI is more immediate and can improve productivity and learning outcomes.

The article then focusses on some fields to appreciate the AI real impacts. It for example has also sparked interest in Art or creative education (Somaini, 2024). While some argue that AI lacks the emotional depth necessary for true creativity, others see it as a tool for expanding artistic possibilities. For example, AI and metaverse technologies are transforming Art education in some institutions, challenging traditional assumptions about online learning in creative fields.

In another important area of use such as foreign language teaching, AI tools such as chatbots and linguistic models offer new possibilities for learners, especially beginners. However, while these tools provide valuable assistance, they often lack the complexity needed for advanced, communicative tasks. AI-driven intelligent tutoring systems (ITS) have the potential to address these challenges, offering adaptive feedback and personalized learning experiences.

However, most current applications remain limited in their ability to fully individualize the learning process. So, despite skepticism about technocentric promises, the potential of AI to personalize learning through data-driven algorithms and learning analytics is significant. However, the reality of institutional resistance could slow its adoption.

Behind myths

Artificial Intelligence (AI in the following article, primarily understood as Generative and narrow AI devices), that is digital communication and computing resources, industrial automation and conversational generative AI, is above all composite. The phenomenon is the result of three joint and interdependent processes:

- the production of technoscientific artifacts (devices based on artificial intelligence)

- The production of scientific discourse (the discipline of artificial intelligence, combining computer science, neuroscience, etc., is still under construction. It hesitates between several denominations and epistemic allegiances.

- The production of myths, imaginaries and discourses around the promises and threats of artificial intelligence, which prophesy and accompany industrial development and socio-political adoption.

Artificial intelligence is an illustration of the society/man/technology nexus in general, where prophecies of “radical revolutions”, whether positive or negative, are commonplace today as they have always been throughout history (Cave & Dihal, 2023). Even more so as literary, mythological (and philosophical) vocabulary and grammar are overabundant, particularly in the West. Let’s recall here the demiurgic and creationist metaphors much used in the science fiction literature of “Judeo-Christian” cultures, comparing man and the creature supposed to imitate him, in an exact mirror image of man’s divine creation out of mud. From the Golem to Frankenstein, from Pinocchio to the many robots of literature and cinema, creatures interact with man, often to supplant him… but generally fail to do so.

In addition to this updated mythological context (Obadia, 2024), contemporary media and culture are saturated with talk of the possible consequences of these technologies. The majority view is that the promises of AI are at least ambivalent and sometimes highly problematic, psychologically or ethically, as when evoking virtual companions or the virtual “resurrection” of lost loved ones, for example. But these observations apply to most of the issues raised by AI:

- the artificial intelligence language models on which today’s conversational AIs are based are highly fallible, make data vulnerable and open the door to cybercrime.

- AI possibly may not be a threat to jobs, but it potentially is to workers and their current skills.

However, the realities are there, and the potential of AI seems (is willing to be) immense, in medicine, industry, administration, education or research, etc.

These multiple debates, which directly affect journalists, academics and artists, are as much a part of the media, artistic and cultural scene as they are of academic reflection, requiring renewed analytical and critical caution, indispensable not to be trapped into myths or ideological blind faith in technological progress and innovation.

A.I. & higher education

Techno-digital sciences and industries, and nowadays A.I., because they affect daily life and working activities are seen in popular culture, industrial-managerial ideologies and science (fictions) narratives as emblematic of a possible new era. ² Let’s recall that it differs strongly from a cultural and linguistic background to another (Cave & Dihal, 2023).

So, technologies, despite they are not what narratives and prophecies still imagine, play an important role. They have a direct impact on contemporary «ways of knowing», that is not only transformation of knowledge (and cultures) that accompanies industrial “third or fourth modernity” (industry 4.0 and 5.0). They also strongly affect conceptions and thinking on training/learning activities.

The construction of legitimate knowledge (i.e., scientific knowledge in our societies) and the circulation/dissemination of representations and ideas are evolving. Education and research, as one of the crucial pillars of those transformation processes are particularly affected and plays a central role, in many ways. Research and Education create and/or accompany the creation of techno-scientific knowledge. They contribute strongly to literate populations to new contexts and to train individual skills (critical?) thinking, etc. for example, techno-digital and A.I. knowledge, as part of scientific Knowledge programs, allows standardization of general knowledge production, through recent development, at a worldwide scale, of academic and research competitions and comparisons. Techno-scientific development is also, and perhaps primarily seen in engineering and administrative standardization, that is generalization of processual and structural tools or devices, and ways of framing professional thinking. Broadly speaking, one can says that the techno-digital and A.I./cognitive sciences « revolution » since around half a century constitute-at least-a change in scale, in capacity to monitor large and complex productive or education systems. It has affected deeply cultural frames and socio-political justifications or discourses. ³

As quite all recent academic articles on AI impact on Higher education point out, Universities are under pressure to create and adopt AI tools in their various functions (Engvall, 2020) and management (research, training, administration, the third mission of innovation and science for or with society), but they are also the driving force behind a radical transformation that has already been announced (Ruano-Borbalan, 2024). The literature on the impact of Artificial Intelligence in higher education, although lacking hindsight and most often based on computer scientists assumptions (Zawacki-Richter, Marín, Bond et al., 2019) or declarative questionnaire for policy-oriented surveys (and therefore highly dependent on the beliefs of those surveyed), converges. AI applications is supposed to have great potential to ‘radically’ change higher education and research. The main aspects highlighted are:

(1) AI’s supposed ability to personalize learning experiences.

(2) The potential automation of administrative tasks.

(3) possible support for research activities.

All these potentialities are supposed to help improve student outcomes and institutional efficiency, conceived mainly in terms of current competitive evaluation models based on ‘quality’ and ‘excellence’. However, most of the analyses also point out that the adoption of A.I. faces major challenges. First and foremost is the need for massive investment in infrastructure and research on and for A.I. The issue of training academic, para-academic and administrative staff is seen as a major problem for the adoption of A.I.-related tools and practices. As we know, there are many other concerns, relating to data confidentiality, the biases of algorithmic operation (hallucinations, linguistic biases in databases, etc.), and the risk of job displacement.

Numerous studies and reports, accompanying public policies or the management orientations of university governance, consider that to exploit the advantages of AI while mitigating its risks, universities need to develop global strategies, integrating their multiple missions. Obviously, since the reference reports used by most analysts come from agencies or universities based in the main countries of techno-scientific innovation, the recommendations are along the lines of competitive public policies in the sector and therefore advocate a massive increase in technological infrastructure (hardware and software), and - above all - training and support for teaching and research staff, and personnel.

To achieve these objectives, in a kind of race and urgency that is often emphasized, Universities (higher education institutions in general) are invited to collaborate directly with industrial partners, both to benefit from the tools and to keep up with developments in A.I., and ultimately to provide students with relevant skills and experience that are directly transferable. It should be remembered that such concepts of strengthening links between industry and education are by no means new, both on a macro-economic level and on a micro-economic level, in the direct link between these firms and vocational training, whether basic or advanced, since the 19th century. It has always been a question of aligning the content and tools of education - essentially vocational education - with technological advances, and at the same time, where possible, with the needs of the commercial and administrative or Industrial world (Troger & Ruano-Borbalan, 2021).

Is AI the key to building “perfect” higher education institutions?

Artificial intelligence (AI) encompasses a wide range of research programs, socio-political discourses, and technological developments that affect teaching, research, management, and policymaking in higher education. However, AI itself does not drive the transformation of these institutions. It is not just a tool like a hammer, but rather a technology that shapes and is shaped by the users through design possibilities (Germonprez et al., 2011). Like other communication technologies, AI must be understood as a “condition of possibilities” that social actors use strategically to reduce pressures and costs in their professional activities. AI is embraced for teaching and research when it eases daily burdens or offers competitive advantages in an increasingly competitive academic landscape. For instance, generative AI tools are now being extensively used by students in systems where evaluation and traditional teaching dominate.

Before exploring how AI affects these activities, it’s essential to recognize the structural realities higher education is facing. Higher education today is vastly different from what it once was (Halasz & Ruano-Borbalan, 2022). One of the most significant changes is the massification of education, with the global student population projected to reach 254 million by 2024 (Altbach & Salmi, 2022). This rapid growth has been accompanied by a widening gap between prestigious research universities and a larger number of diversified institutions. The top-tier universities, often international and digitized, set global standards, while others remain more locally focused.

Qualitative changes are also evident. Increased student mobility from developing countries to established knowledge centers, the dominance of English as the academic lingua franca, and the growth of technoscience and digital tools have reshaped both teaching and research. Furthermore, university management has shifted toward a more solution-oriented, professional model driven by managers and engineers rather than traditional faculty, emphasizing innovation and efficiency (Ruano-Borbalan, 2023).

The global landscape of higher education has thus become both unified and polarized. Prestigious research universities, particularly in the U.S., Europe, and Asia, set the standards for academic excellence, while professional institutions focus on education and research with local applications (Brink, 2018; Popp-Berman & Paradeise, 2016). These shifts have transformed admissions criteria, deepened inequalities, and influenced the definition of knowledge itself. Additionally, universities have increasingly adopted an entrepreneurial model, competing for resources, students, and research recognition in an international marketplace (Brint, 2018).

National policies also play a role in shaping higher education, with countries such as Japan and China establishing centers of excellence and European nations following suit through initiatives like Erasmus and Horizon Europe (Croucher, Wang, & Yang, 2023). These programs emphasize multidisciplinary approaches, partnerships, and innovation to address economic and technological challenges (Ruano-Borbalan, 2019b). Yet, the drive for innovation in universities, spurred by the global expansion of business and engineering schools, is rooted in neoliberal principles and technoscience (Ruano-Borbalan, 2019a).

In this context, the role of universities has expanded beyond teaching and research to fostering industrial and territorial innovation, as well as cultural and societal engagement (Unger & Polt, 2017). However, these changes have met with resistance, particularly in countries like France, where reforms aimed at aligning with European standards have sparked tensions between tradition and new performance-driven models.

The transformation of academic careers is another area where tensions are visible. The traditional “Humboldtian” vision of the academic profession—centered on the union of teaching and research—is increasingly challenged by the rise of evaluation systems and administrative responsibilities (Redding & Crump, 2019). Today, academic careers are fragmented into specialized roles, with increasing differences between those who focus on research and those involved primarily in teaching or administrative duties. This fragmentation of academic identity has led to a concentration of careers on individual achievements and a growing divide between professional and academic spheres (Endrizzi, 2017).

As AI tools become more integrated into higher education, their adoption will be driven by their ability to ease competition and reduce the workload in teaching and research. For example, generative AI is already assisting students and researchers in producing work more efficiently. However, the promise of AI revolutionizing individualized learning remains largely unfulfilled. Just like previous technological revolutions, the impact of AI on education is complex, varied, and often incremental. AI alone will not disrupt or transform education; its effects will be shaped by broader institutional and professional changes.

Will technology disrupt higher education?

The hypothesis that technology define social or institutional change is an old belief, often refueled, and particularly present in our societies during the industrial period. Despite its strong presence in many analysis, it is not true. Not because technology doesn’t plays very important roles, but simply because there is not only one factor to understand societies change, as contemporary historians and social scientists shows: even if there are obviously systemic links between economic growth and technological development (Mokyr, 2005).

Considering Education we can take only one example, and recall that in 1913, Thomas Edison predicted that cinema would replace books in education within a decade. Ten years later, he admitted that this had not happened, but he believed that it was still inevitable. We now know that it was not true, and more generally we know that communication technologies and technologies for creating and disseminating knowledge such AI, even intellectual knowledge, do not replace previous technologies, but rather add to them. They add complexity to the systems that enable humans to remain what they are: first and foremost, political and social animals for whom the socio-political link through the exchange of communication is an evolutionary process, and therefore crucial skills and competencies. This tendency to ‘over-predict’ technological revolutions has persisted, particularly in complex areas such as education, culture, economics and politics. Let’s not forget that they constitute conceptual tools for commercial or political promotion, aimed at accelerating the purchase or adoption of products and visions, without the authors of such speeches automatically being in bad faith.

The focus today, of course, is on artificial intelligence (AI), which has advanced rapidly thanks to the joint development of AI programming, digitalization, the Web, robotization and interconnected scientific programs, including cognitive science, linguistic, anthropology, etc. As one perfectly knows, a major ‘breakthrough’ occurred in the 2010s, marked by huge progresses in ultra-fast information processing and hyper-massive data categorization. Recent conversational AI tools, such as ChatGPT, have rapidly gained in popularity, improving a variety of standard tasks.

The development of AI is being driven by powerful industries and considerable public and private investment, particularly in the US and China. AI is expected to have a significant impact on the global economy, national security and international trade, as well as on social and professional practices, particularly higher education and research. However, the field is controversial, with concerns about data bias, misinformation and the ethical implications of the widespread use of AI. Despite its potential, AI often reinforces existing societal inequalities and enables increased surveillance and control.

The dominant approach to AI, focused on data, speed and processing power, was not inevitable and has not always been so data centric (Ander, 2023). The rapid advances of the 2010s have led to a ubiquitous digital and communications environment that both empowers and controls individuals, particularly in the urban middle and upper classes. This dual role of technology - empowerment and control - reflects the wider impact of AI on contemporary society.

At the dawn of the 2010s, new techno-centric - and often industrially-involved - prophets announced that education, too, would change dramatically, thanks to the capabilities of mass distance learning. The two main promises were that learning would be individualized (thanks to Moocs, for example) and that almost every human being on earth would be able to benefit from the “best” educators and professors from the elite Western universities. Here again, the revolution has not taken place let’s simply recall that Moocs are appallingly poor in terms of cooperation: their enrolment has constantly decreased, and their overall completion rate is very low between 5 to 12% of registrants (Jordan, 2015; Reich, 2020).

Educational systems, like human learning, are so complex, with such diverse aims and constraints, that rapid global improvement seems out of reach. Education is just one of many areas in which the political, anthropological or social revolution promised by technology is more complex than techno-optimists often suggest. If new forms of communication and virtual or immaterial cultural practices have emerged and spread, they have done so within the confines of pre-existing consumer habits, where social and political relations continue to retain their intrinsically “human” nature.

Today’s technological advances, including those enabling complete individualization of educational paths, cannot replace the need for human intervention. At their best, these technologies enable individuals who have undergone a formal, human education to trust their ability to learn independently, to deepen their knowledge. However, this process tends to exacerbate inequalities in access to education, since it assumes already acquired critical thinking and problem-solving skills (Reich, 2020).

We can add to this very important understanding of Technologies and Education entanglements, the increasing complexity of these systems, and clearly the development of A.I.-learning environments, that perhaps will allow education to better adapt learners’ heterogeneity, possibly enabling high-level learning and practice, but is just in the first steps of its development. Taking real education life practices, we really need to be cautious about the speed of change in educational practices that the technology is likely to generate. To take just one example, with regard to languages teaching, we note that while foreign language learning programs are among the most widely used applications on the Internet, and computer-assisted language learning (CALL) software has been widely used for several decades, it is no less true that until today, only a small number of applications offer a learning experience that goes beyond linear pathways, simple feedback routines (good or bad) or a “one-size-fits-all” approach (Schmidt & Strasser, 2022).

AI and teaching/learning processes and realities

As said above, Generative AI, as part of a broader digital and cultural ecosystem, is potentially reshaping our relationship to knowledge, influencing our lifestyles, and transforming our aspirations. Given the rapid advancement of AI tools in professional activities or social interactions and its growing market potential, the intersection between AI and education has expanded dramatically. AI tools are becoming increasingly sophisticated ⁴ and could be especially useful to beginner learners across many disciplines, depending strongly to the capacity of adapting and configuring learning spaces and environments to fit human learning processes. They can offer user-friendly interfaces and are accessible on both mobile and desktop devices, allowing for flexible use across different times and spaces. With their multimedia capabilities, these tools could enable learners to engage in exercises using audio, video, and written formats, particularly benefiting subjects like language learning, mathematics, computer science, and disciplines that rely on memorization or gestures. Furthermore, students may find comfort in these anonymous, AI-driven scenarios, where their mistakes or personal details remain private, and paralinguistic cues of judgment or.

The recent advancements in data-driven algorithms and multi-layer technologies, in combination with sophisticated analytical processes like learning analytics, offers the promise of personalized learning, formative assessment, and student-centered educational experiences. Educational data mining could, in theory, enable highly individualized learning processes. However, while this potential exists, the reality of institutional and pedagogical resistance within educational organizations should not be overlooked (Reich, 2020).

Once again, technologies are not capable alone, without insertion in appropriate learning environments and spaces, to individualize education. If competitive exams, hierarchization through assessments, and permanent negative reinforcement are to remain, and there is no real reason for a major change, technologies will not modify the very educational processes, which are not only to make student learn!

But learning spaces and environment are changing, even if the socio-economic and socio-political functions (to distribute people in industry and administration, selecting elites, etc.) AI technologies have the potential to contribute blurring the boundaries between formally structured learning environments, such as schools and universities, and informal learning opportunities, that occur during leisure time or at home. This trend is not new, but it is becoming more prevalent and accessible for learners. However, higher education systems, including teaching, evaluation, and the conferral of diplomas, may not fully embrace these changes, except perhaps in some specific parts of Higher Education such research training (Masters and PhD) and research, where the impact of AI could be considerable, primarily because it affects immediately the productive capacity, reducing for individuals the energy costs for high value resources such information or time consuming writing activities for example, in their anxious and highly competitive professional life.

As AI continues to evolve, its integration into educational environments will likely play an increasingly significant, but differentiated, role in shaping the future of teaching and learning, depending on sectors, disciplines, kind of pedagogy or evaluation used, etc. Those tools offering both challenges and opportunities for educators, learners, and institutions alike. The best to measure it is to take some learning and teaching fields.

A.I. in art or creative education

There is for example, a huge interest for AI in Art or creative Education, both for practices and for more philosophical questions (Zhao et al., 2024). In this last respect, impact of digital tools and artificial intelligence (AI) has sparked a significant debate among artists, designers, architects, computer scientists, and humanities scholars. Central to this discussion is the question of whether machines can genuinely be creative (Du Sotoy, 2019). While this issue may seem less urgent compared to concerns about AI’s role in the workforce, autonomous weapons, or the potential for AI to develop consciousness, it touches upon the fundamental nature of artistic and creative practices.

Creativity extends beyond the mere production of aesthetically pleasing works; it involves the generation of new ideas, strategies, and perspectives, all deeply intertwined with human cognition and emotion. In traditional philosophical thought, creativity is considered a hallmark of human experience. This brings into focus profound questions about the essence of creativity itself as we contemplate AI’s role in art and design.

A major argument against AI’s ability to be genuinely creative lies in the relationship between art learning or practices and emotion. Human creativity is still more often supposed to be closely linked to consciousness and emotional experiences, that A.I. would not be able to reach. ⁵ When we create art, we infuse it with aspects of our inner emotional world, making it a reflection of our subjective experiences. This process is not merely technical; it is an expression of consciousness that machines, lacking subjective experience, cannot replicate. Machines do not “feel” or “understand” in the way humans do, which is why many argue that while AI can produce art, it lacks the depth and authenticity derived from human emotional expression, leading some to view such creations as artificial or inauthentic…for now (Andler, 2023; Du Sotoy, 2019).

Despite this, contemporary artistic practices increasingly integrate digital and AI tools, reshaping traditional forms of artistic expression. ⁶ In fields such as design, visual arts, crafts, video games, cinema, live performance, and artistic mediation, AI is being used not only to enhance creative processes but also to transform them entirely. These technologies offer new possibilities for expression and experimentation, pushing the boundaries of what can be achieved in art and design.

Moreover, digital tools and AI are significantly altering social, educational, and training practices within the arts. (They are transforming how artists learn, create, and share their work, and they are influencing the evolving roles of the arts in society.

A pertinent example of this transformation can be seen in the evolution of a Chinese private preparatory school, training students for competitive entrance examinations for French art schools. ⁷ Established in 2014, this school initially offered traditional art education workshops, combining both in-person and distance learning formats for the presentation of art pieces. However, following the pandemic, the program underwent a complete transformation, moving entirely online and integrating AI and metaverse technologies.

The online art education platform is structured around three pedagogical strands: language and culture, art and creation, and art theory with critical thinking. This initiative seeks to fully transpose pre-pandemic educational modules into a distance-learning format, surpassing the temporary measures adopted during the pandemic, when only art creation courses were held online. Each educational component has its own organizational particularities, yet together they aim to provide a holistic learning experience.

One of the initial findings from this transformation (the case is under analyses in a PhD process, not yet published), though obvious from an economic and management perspective, is rather counterintuitive from a learning standpoint. The complete shift to distance learning, with the integration of AI and «metaverse» immersive tools, does not seem to negatively impact the performance of Chinese students in competitive examinations for French art schools. This challenges assumptions about the limitations of online and AI-augmented education in fields that rely heavily on creativity and artistic practice, and even in all learning processes where the «doxa» is to promote so-called Hybrid or blended forms of education.

Finally, we can assume that the integration of AI and digital tools into art and education is reshaping both the creative process and the pedagogical models, raising complex questions about Art Education and the future of creativity, human emotional expression, and the evolving role of machines in artistic collaboration.

Foreign language teaching

In foreign language teaching also, the rise of large language models (LLMs) and chatbots has significantly fueled the hypothesis of a technological transformation (Schmidt & Strasser, 2022). These tools, with access to extensive linguistic corpora, have become increasingly sophisticated, offering substantial potential for language learners, particularly beginners. Given that basic words and phrases often follow clear structures with minimal semantic, syntactic, or lexical complexity, such tools provide valuable assistance during the early stages of learning. Designed with user-friendliness in mind, they present intuitive interfaces accessible ubiquitously across mobile and desktop platforms. Additionally, their multimedia components enable learners to engage in exercises that incorporate audio, video, and images while practicing through speech recognition or written tasks. For anxious learners, these tools offer relatively anonymous interactions, allowing them to practice without the fear of face-to-face judgment or the paralinguistic cues that might indicate disinterest or criticism.

Modern foreign language teaching, however, places complex communicative learning tasks and task-based language learning (TBLL) at its core, an approach that emphasizes the integrated development of functional communicative skills. Effective language learning tasks should promote the use of the target language for meaningful communication, focusing on message content and meaning while maintaining authenticity. These tasks are designed to culminate in well-defined linguistic outcomes. At advanced levels, learners are expected to communicate and act authentically in real-world scenarios, as this type of production is crucial for language acquisition according to the production hypothesis.

In fostering intelligent foreign language practice, priority must be given to learners’ communicative needs, particularly fluency and competence in the target language. Practices should remain authentic in both language and content while promoting autonomy and individualized feedback, encompassing linguistic corrections and strategic communication approaches. Furthermore, language practice should be aligned with cognitive and affective learner needs, goal-oriented, and transparent, focusing on the development of metacognitive and linguistic skills. This marks a clear departure from older methods that relied on rote memorization of language structures devoid of communicative purpose. By tailoring these processes to individual learner needs, foreign language instruction aligns with competence-based tasks, ensuring that learners are engaged with exercises stimulating enough to match their zone of proximal development (ZPD), as theorized by Vygotsky. High-quality feedback and social interaction with peers are essential to this process, as they contribute to the learner’s development within a supportive environment.

A fundamental question arises as to whether, and to what extent, digital learning programs can adequately support these learning objectives. AI-driven intelligent tutoring systems (ITS) have the theoretical potential to address several of the central challenges posed by individualized and adaptive learning environments. However, the effectiveness of existing language learning programs remains in question. Developing intelligent language tutoring systems (ILTS) involves significant challenges, particularly in establishing an accurate automatic analysis of learners’ language development, known as interlanguage. Natural Language Processing (NLP) technologies must account for the vast variability of human language, especially in the context of learners’ evolving linguistic usage. Moreover, this modeling must be integrated with both task and learner modeling, using learner corpora to analyze the effects of various tasks and exercises.

While computer-aided instruction (CAI) and ILTS have the potential for interactive problem-solving, intelligent error analysis, adaptive feedback, and personalized curriculum sequencing, such systems are still relatively scarce. Many current applications provide poor-quality feedback, focusing on communicative objectives or isolated grammar and vocabulary exercises. Furthermore, the sequencing of these exercises tends to be rigid and predetermined, lacking the flexibility required for genuine individualization. Most current systems do not yet incorporate the artificial intelligence methods necessary for adaptive learning. As a result, while the market is saturated with computer-assisted language learning (CALL) programs, truly intelligent computer-assisted language learning (ICALL) remains quite limited.

For teachers and trainers to effectively support learners, they need detailed insights into the linguistic requirements of exercises, learners’ individual strengths and weaknesses, their errors, affective attributes, and overall progress. Developing an intelligent computer-based system capable of supporting adaptive foreign language learning requires an architectural structure that includes specific interrelated components. A domain model provides information about the language being taught, along with exercises of varying difficulty, empirically validated against grammatical concepts. An assessment model continuously measures learner performance, identifies error types, monitors time spent on learning, and tracks the use of feedback and tutorials. Meanwhile, a learner model updates and collects data on learners’ mastery of content as they progress, guiding personalized learning experiences. This architecture, directing the entire learning process and determining appropriate responses (e.g., feedback, new content, revision), must be deeply integrated with learning analytics, NLP, and error analysis. The task of designing such systems, which combines diverse technological and pedagogical demands, is non-trivial, requiring interdisciplinary collaboration among experts to fully realize the potential of AI in language education.

Conclusion

If we want to imagine the real impact on teaching/learning processes, let’s have a kind of imaginary description of what a classroom could be in a few decades, in countries and places where resources are abundant: the learning space and environment/classroom will be an evidence-based blended learning environment, with bring-your-own-device solutions in every classroom and subject, high-speed Internet connections, and standardized content management systems, hybrid coursebooks and learning materials forming the backbone of the infrastructure. Learners from different backgrounds will have equal opportunities to participate. At this time, the printed textbook will have become an interactive, multimedia and adaptive learning and practice environment, suited to the needs of face-to-face teaching and phases of individual practice and self-directed learning. Experts in the disciplines-based knowledge, pedagogy, computational education & A.I., machine learning and Big Data, robotics, learning psychology, and cognitive sciences, as well as multimedia designers, should work closely together to create multi-learning environment. The future digitally enhanced learning place and environment will help diagnose learning needs and progress, offering direct access to differentiated, needs-based support services. The learning platform deployed will therefore be an AI-based digital learning support system, a resource, a tool for students and teachers, which will be used to add value to the learning process. The classroom of the future will cleverly combine the benefits of digital learning with proven computer-free methods.

The classroom or learning platforms of the future should combine the benefits of digital learning with proven methods, content and tasks (without computers for face-to-face teaching), which will remain indispensable and highly significant for successful learning. Even in this possible future, education will need teachers: well-trained, data-literate, competent, critical and reflective in their use of media and technological support. Educators will use empirically established digital scenarios based on useful learning content and linked to meaningfully constructed and stimulating learning tasks and exercise opportunities designed to support individual learning. A sustainable approach to familiarizing teachers with AI technology as part of a non-technocratic narrative would likely involve a hybrid transition period until the point is reached where people feel confident in replacing a truer assessment of an individual’s learning and strengths.

But whatever will be the future, Higher Education faculties and administration are facing important challenges, including the importance of a promoting new literacy (“digital literacy”, “data literacy”, “AI literacy”) that enables the various players to adopt an enlightened attitude. Indeed, to accurately prepare future of education based on nurturing human capacity of reasoning, develop creativity and critical thinking, and avoid pitfall A/I tools can and surely will enhance, we need to clarify the terms used (data, digital, generative AI, etc.), and place the issue of AI in education and higher education within an already long history of attempts to introduce technology into learning in the context of the “industrialization of training”.

A renewed humanism, -imposed by techno-digital and A.I. devices impact on Education, research and society-, constitutes the very heart of Higher Education challenges, and research practices. Such concerns are precisely the kind of issues that need to be addressed in terms of learning/teaching and production of knowledge within Higher Education systems. It’s precisely in line with the recommendations made by global public policy players (UNESCO, European Union, International Science Council, Academies, research agencies, etc.) to use and implement new technoscience and A.I. tools, but also at the same time promote cross-fertilization between science, art, humanities, behavioral and social sciences. It’s not just a question of tirelessly promoting the use of digital technologies and artificial intelligence. It’s about understanding and thinking about the interaction and growing consubstantiality of human systems and technologies in the production and dissemination of knowledge.