Introduction

This paper approaches and evaluates the new role which Artificial Intelligence (AI) has assumed in the last few years with respect to advanced, academic education focusing on the issue of creative and critical writing. To assess solidly what AI might really mean in practical terms, I will first reflect on technology and pedagogy at large, then consider previous innovations that have affected education, to turn subsequently to the specific concerns of AI and academic writing. Based on practical experiences and experiments and on theoretical reflections, the argument points out the considerable dangers of AI for the intellectual development of students who really have to work on their own to develop critical thinking. This is not to say that we should reject AI altogether since we all have used it already in many different fashions whether we are familiar with AI or not. We are in a very fluid environment currently, and there might be many areas in advanced education where this new technology might be useful. But before we proceed too quickly, it is most pertinent to consider what specific impact AI has had on our students so far in a narrow teaching context.

Materials and Methods

Paradigm shifts have always moved humanity forward, which does not necessarily entail automatic progress. But changes happen all the time, both in our teaching material and increasingly teaching technology. COVID-19 created or enforced a massive re-orientation, and most of us lived up to the challenges and actually discovered the considerable advantages of teaching online via zoom, Microsoft Teams, or other internet media because it made it possible for many more students to join our classes irrespective of the geographic distances. Distant learning has become a reality that works quite well and does not present a dystopian vision of utter alienation and loss of interest and motivation. The initial shock when new technologies enter our educational realms tends to be significant, but in the course of time, with much practice, many adjustments, and sorting out what is practical and what is not, the drama associated with those new teaching tools often disappears.

Digitization of texts or images has also implied that many more resources have become easily available, especially older texts, such as from the nineteenth, sixteenth, or thirteen centuries. Manuscripts can be viewed nowadays much more intensively via the digital version than in the past when it was only possible to visit the respective library or archive and then to ask for the permission to inspect a specific manuscript, which might not be possible because it is under repair, loaned to an exhibition, or in a too fragile shape. The zoom function of digital copies allows us today to examine those documents much more closely than ever before, that is, even more in-depth than the naked eye could do in the past (see, for instance, van Lit 2019; van Renswoude 2024).

However, the real issue in higher education is not so much the access to primary documents or their availability, but the perennial problem of plagiarism, that is, how students engage with them, if at all, and whether they learn how to establish their own intellectual position vis-à-vis a poet’s, writer’s, artist’s, or philosopher’s statement/s. We have mastered that issue mostly by means of sophisticated software programs such as Turnitin, which have developed astounding capabilities of identifying specifically the sources that a plagiarist might have used. When someone copies illegally from a published or unpublished source, it is by now fairly easy to identify where the original text has been lifted from, even when the source was another student’s paper.

Results

As we all know by now, by contrast, AI-generated texts operate very differently and do not simply copy any published work, such as Wikipedia or any digitized article or book-length study. Instead, AI creates truly new texts by culling from a wide array of sources various narrative elements and combining them in a surprisingly meaningful manner, at least on the surface. The outcome then proves to be a ‘new’ text, which makes it virtually impossible to identify where the author has drawn his/her material from. The teacher or professor is then at a loss of how to evaluate the submitted product, unless it can be rejected outright as a case of ‘plagiarism,’ which undermines the entire concept of grading, which thus hurts both the entire class and the instructor, but eventually also the person who had resorted to AI for help.

Technological Challenges

A very recent example of the problems involved in AI-generated texts pertains to news or news alerts sent to some iPhone users. Apple has now realized that its product, the beta version iOS 18.3, has tended to produce fake news without checking tools implemented. As reported by the Associated Press, “Apple said it is disabling the AI-generated feature for news and entertainment while it tries to fix the issue causing the technology to fabricate information – a problem often described as ‘Hallucinations’ within the industry” (Apple statement online, link 1). We also know that Meta, formerly Facebook, recently suspended fact-checking, replacing it with a community-driven system that opens all doors for malicious misinformation, which in turn destroys the reliability of this and other social platforms (Meta statement online, link 2). Whether we talk about plagiarism or fake news, the critical issue remains the same, and this then allows us to reflect further on the use of unacceptable AI-generated texts submitted by students for grade. The danger for our society at large is obvious, and it is prone to get worse than better in the near future. We in the teaching trenches, however, have the great opportunity to work with the new generation of students to prepare them for those egregious challenges of how to determine what is a fact and what is a fake; how to read texts critically and to respond to them independently. If we do not manage to secure this intellectual capacity in each individual, then our free and democratic society will be severely challenged.

Other big issues concerning the use of AI, though not pertinent to the specific focus of the present paper, consist of the amount of data collected by AI and the question where that data is stored, who can have access to it, and whether security risks are properly taken into account. Then, we also have to consider the environmental impact of AI which requires enormous amounts of energy to operate (links for websites 3 and 4; both last accessed on Jan. 22, 2025). Finally, AI seems to solve many problems, ease production, improve agricultural yields, and accelerate operations, especially by making them more precise or accurate (trains, airplanes, commuter traffic, all kinds of data-collection operations, large-scale productions, consumer analysis, etc.).

However, the big question then quickly emerges who profits from this technology. In agriculture, for instance, AI is too expensive for most farmers across the world, so it serves primarily big agro-businesses. In the field of medicine, however, when the researcher or practicing doctor faces big data, AI promises to be a most welcome help, as long at it does not replace the doctor as the human factor in all healthcare issues. We face, in other words, still many serious questions regarding AI, or we might only be at the beginning of a technological revolution at the end of which humans will lose out to robots or develop meaningful partnerships. No wonder that AI challenges education as well, the topic of this paper (see the presentations offered at the 3rd Global Summit on Artificial Intelligence, HSPI Open Access Publisher, London, Jan. 22, 2025: Scientific Program for 3rd Global Summit on Artificial Intelligence (GSAI).pdf).

The Individual Student and AI

In essence, the individual who resorts to AI for whatever reason is obviously driven by several motivations. AI makes it easy to produce an impressive result when a student does not have the time to do so, and it promises to create a written work that seems sophisticated enough to get a good, if not the highest possible grade. On the surface, the end-product meets all the expectations, and in the future, the person who has utilized the computer as a tool will thus believe that s/he can do so in the future as well, and thus becomes entirely dependent on AI for all subsequent tasks that involve the establishment of a unique narrative with a solid thesis, argument, and conclusion. It seems highly questionable, however, whether this person would have understood in the first place why there had been an assignment to produce, for instance, an essay or a thesis paper. Many current scholars tend to welcome AI even in writing classes (cf. the contributions to Kersten and Ludwig, ed., 2024), but, as my examples and reflections below will indicate, there are more problems and serious challenges in that matter than constructive solutions, at least currently. Writing itself is a difficult subject matter and requires a lot of practice over a long period of time. Resorting to computer help might be valid and useful when it concerns the mechanical aspects, such as spelling or grammar. However, the ultimate purpose of any writing consists of developing and expressing one’s own thoughts and to gain an independent position within a wider discourse.

Three metaphors have been developed that shed a lot of light on the problems with resorting to AI in a learning situation. When someone builds a robot to do physical exercise, then there will not be any result for the user because the machine has taken over the very activity through which muscles can be built. Similarly, if someone uses a robot to play perfect piano music, s/he might enjoy the music but will not develop the skills to perform that music him/herself. In the field of foreign language learning, the situation is particularly problematic because there are, indeed, a number of great advantages with AI when people have to communicate with each other without having any knowledge of the other person’s linguistic code. How good that form of communication might be, however, remains a very debatable issue.

When an Iranian and a Russian, for instance, have some business deals with each other and yet cannot understand the respective other language, robotic translations can solve many practical issues, though the danger remains that most of the nuances, cultural sensitivities, personal preferences, and individual needs are not taken into consideration by the machine. After all, human communication is not a one-way street, it is often characterized by ambiguity, nuances, emotions, and a variety of conflicting perspectives, desires, and needs. Many times, critical words such as ‘freedom,’ ‘independence,’ ‘responsibility,’ ‘spirituality,’ or ‘individuality’ can carry many different meanings and present serious problems for translators, both in historical (vertical) and in global terms (horizontal). ‘Democracy,’ for instance, can stand for very different political concepts, depending on the circumstances and the specific discourse. Hence, the individual speaker would rely on a machine that might or might not render his/her ideas properly into another language, which might, of course, become a moot point in the near future, though I seriously doubt that we will easily reach such a point, if ever.

Nevertheless, as everyone knows who has ever worked with words, sentences, poetry, or texts, the subtleties of human expressions are, while often perturbing, perplexing, and even unsettling, a huge challenge and yet also the very beauty and potentiality of our language which, I dare say, no machine can ever replicate because it does not want to do so being programmed to be precise, correct, and unambiguous. As we all know, however, pure logic and rationality, as desirable as they might be, are not the basic components human existence is made of. Randomness, arbitrary decisions, irrational operations, and simply quirkiness constitute some of the very essentials of our lives, in religion, the arts, literature, philosophy, ethics, morality, and other fields of human endeavor

AI in Foreign Language Learning

Without even any sensitivity regarding the character of a foreign language, cultural differences become intractable and can easily lead to extensive miscommunication. After all, we can hardly expect that most of our students can quickly develop language skills at the level of native or near-native speakers. However, having been exposed to the challenge of learning a foreign language without the help of a robot creates already the fundamental degree of sensitivity toward languages as such, which includes the own language as a construct that transforms and develops all the time. A person who learns a foreign language really learns in that process much about the own language. As ideal as it might be to acquire a high degree of linguistic competence in a foreign language, the realization itself of all the difficulties involved and the own mistakes in speaking a foreign language can create a sense of respect for foreign languages as such (see the contributions to Vanderheiden and Mayer, ed., 2020).

AI offers many technological advantages, but it also deprives the individual of the human experience with languages, the constant exposure to trial and error, and the joy of accomplishing a modicum of communicative skills and achievement. It does not need to be stated explicitly and yet deserves to be highlighted here once again that human life is deeply determined by language, and the more we as people engage with words, sentences, and poetry, the more we can discover ourselves as social individuals dependent on our community determined by a commonly shared language, value system, and political ideals. Language is constantly evolving, and we as people contribute to that process and profit or suffer from it as well. We educate students at the university level so that they become aware of these issues and are more or less in charge of their own language and hence their culture and identity. that is, of their learning process and their intellectual growth as individuals.

Of course, we have already become very accustomed to the assistance by language robots such as Grammarly that helps us to detect grammatical or spelling errors while we type. But the difference to an AI-generated text and one which has been polished and corrected by the computer is vast. Everyone wants to produce texts that comply with the norms, are free of errors, and appeal to the reader. Both in the past and today, many writers resort to friends, colleagues, or copy editors to look over their texts, since we all know that two pairs of eyes see more than one pair only. Worse, we as readers of our own texts tend to ignore even glaring mistakes because in our common reading processes we do not read letter by letter to form or sound out a word but we identify words holistically and can thus speed up our reading of the text, often even diagonally across the page. Naturally, this makes us somewhat blind to our writing because we are so familiar with it and read what we want to see on the page, even if that is not the case.

Of course, the opposite should be the case, but we easily lose the ability to read critically, carefully, and closely. Some of the strategies to break up that familiarity consist of using a different font type, to increase the font size, or to read the text from back to front. We also know only too well that resting our eyes, i.e., the text, and then return to it refreshed allows us often to see the text through new lenses, which allows us commonly to do self-corrections. Grammarly and other software programs do the same for us, only faster and more consistently and they enhance our own ability to improve our writing skills. However, it still remains our own writing, and there are no serious conflicts with authenticity.

Discussion

AI and Academic Writing

AI-generated texts constitute a very different ball game and endanger or impoverish us humans in our ability to formulate clear ideas, to think through a problem, to solve issues, and to express our concerns and concepts in our own terms, for ourselves. The true issue with AI seems to be that it produces a deceptively elegant and cogent argument based on a variety of opinions that are smartly balanced with each other by way of linguistic connectors or subordinate conjunctions that seemingly convey a sense of complex knowledge that could be there or that could be nothing but a soap bubble: while, whereas, although, on the other hand, seen from a different perspective, in contrast to, and the like. In scholarly writing, we expect the implementation of principles of honesty, justice, respect, and balance to reach valid conclusions based on the solid arguments developed by means of rationality and facts as far as a critical reading can establish.

All scholarship thrives on the the concepts of falsification and verification, as Peter Abelard (d. 1142) had already developed it in the twelfth century (see the contributions to Yu and Gracia, ed., 2003; Bezner 2005, Binini 2022). Of course, he could not have foreseen the emergence of AI eight hundred years later, but through his development of the method of sic et non, i.e., a rational examination of the Scriptures and other holy texts, he had laid the foundation for our own critical assessment of every aspect in scholarship (Abelard 2008). Could it thus be that this venerable foundation of critical thinking might now be replaced by a robotic process determined primarily by mechanistic operations without any ethics, morality, a sense of honor, and the dignity of human rationality? (Marenbon 1997).

AI has triggered enormous discussions around the world, but since we are currently in the middle of the paradigm shift (or only at the beginning), the opinions about its value, usefulness, or dangers are multiplying on a daily basis (see the contributions to Murugan, Periasamy. and Abirami, ed., 2025, who outline numerous scenarios of a productive integration of AI in higher education, such as in the STEM fields, in self-learning situations, and tutoring settings). For Bowen and Watson (2024), there are numerous advantages with AI because we only need to learn how to harness the tools effectively when we turn to AI. According to the online book content, we are told,

How AI is revolutionizing the future of learning and how educators can adapt to this new era of human thinking. Artificial Intelligence (AI) is revolutionizing the way we learn, work, and think. Its integration into classrooms and workplaces is already underway, impacting and challenging ideas about creativity, authorship, and education. . . . José Antonio Bowen and C. Edward Watson present emerging and powerful research on the seismic changes AI is already creating in schools and the workplace, providing invaluable insights into what AI can accomplish in the classroom and beyond. By learning how to use new AI tools and resources, educators will gain the confidence to navigate the challenges and seize the opportunities presented by AI.

We can certainly agree with them that “In the age of AI, critical thinking skills, information literacy, and a liberal arts education are more important than ever.” But can we achieve these goals by simply embracing a robotic system that takes over our own thinking and produces what we are supposed to create by ourselves? What constitutes cheating, plagiarism, false pretenses, and the like, especially within the field of the Humanities? As Pang (2024) now admits, the use of AI in Foreign Language courses can constitute serious problems of ethical and pragmatic kinds because when learning itself is no longer necessary, then the human brain is in danger of shutting off and relying on a machine to take over the task of thinking and all mental creativity. This implies, of course, following Pang, that we discriminate as much as possible in assessing what AI can do well and what AI should not be used for if we don’t want to undermine the entire educational process (see also Chan 2024) and hence the process of identity formation.

As with all new technologies that impact human life, take over human tasks, ease human labor, and intensify productivity and profit at least for the owners of the machines, the influence of AI on all of us carries heavy ethical implications, especially because most teachers or instructors across the globe have to make their own individual decisions regarding how to evaluate AI-generated texts, what to expect from their students, and how to guide them toward the goal of independent thinkers and critical writers. We might gain more clarity in this issue in ten to fifteen years, but currently, the range of approaches, attitudes, perspectives, opinions, and methods is a myriad (Bayer and Grimme, ed., 2025), with many different interests at stake that often contradict each other.

As in the case of most previous technological inventions, concerning AI, there is no effective way of resisting its development and further implementation. We have virtually no real choice and cannot stem the flood of AI-generated products also in higher education, and hence its impact on us in our daily lives. Yet, using AI for data collection, for instance, also within a literary or narrative context, has proven to be rather effective (Martin 2025, talks about a ‘technologically mediated landscape’ we all live in already). But, as in the past when the computer and then the internet entered the picture, we have the opportunity and face also the ethical obligation to examine it critically and, under certain circumstances, to allow it to gain its own position within the broad range of methods and tools in pedagogy (see the comments by Aaron, Abbate, and Allain 2024).

The Obvious Dangers with AI

In my experience so far, however, AI has been a fundamentally dangerous tool particularly for young students engaged with reading and writing who do not yet know what they should study, why they study what, and how to go about studying in a serious fashion so that they can get the most out of that experience for themselves. What I want to contribute here does not have much to do with the large picture of AI in global terms. Instead, the question to be addressed subsequently pertains to the common task we as instructors ask our students to do in many, if not most of our classes in the field of the Humanities, i.e., to write papers. There are many ways of getting students engaged in writing about a subject matter, such as in the traditional format of a fully-developed and fleshed out paper consisting of a thesis, argument, and conclusion, which is well vetted and solidly supported by relevant scholarship, just as in this paper that I am presenting here. There are also numerous technological learning management systems such as Top Hat that facilitate much writing in the classroom at regular or irregular intervals, often in response to a critical issue (Classen 2020; see also the official website of Top Hat at https://tophat.com/). In that case, since many of the topics raised could be formulated ad hoc, and since the time to answer would be limited, there seems to be a minimum of danger that a student would quickly turn to AI to get a more polished and sophisticated answer on the spur of the moment.

Creation of Critical Papers

The real issue emerges when students have to compose and submit papers in an academic manner that require critical acumen and basic research skills, both geared toward the development of creativity and independence. Given that those papers must be produced at home or in the library, this provides students with much time to draw from online sources (committing plagiarism) or to ask for AI support writing the entire paper or at least sections of it. To my own surprise, whenever that happened, it proved to be very easy to determine that the paper was composed by a machine and not by a human being. The consequences were rather dire for those students whose ‘cheating’ or plagiarism was detected since they received either a 0 for the paper or an F for the entire course.

The difficulty for the instructor rests, of course, in proving that a paper was AI-generated. Astonishingly, although I have had already a good number of paper submissions that my graders/teaching assistants and I could recognize as such, none of the students charged with this form of ‘plagiarism’ defended themselves or rejected the accusation. Most of them admitted their fault by their silence. By contrast, in several cases, students openly admitted that they had asked a computer (ChatGPT et al.) to do the work for them and that they felt ashamed now that they had been caught. In those cases, we had a very good teaching situation which then solved the entire issue peacefully and led to the students’ considerable growth in ethics and academic standards.

All forms of plagiarism can be turned around into a good learning experience if the involved student lives up to his/her mistake and understands fully why and how his/her method was wrong and dangerous concerning his/her academic integrity and individual honor and dignity. Of course, all this would make sense only if the purpose of writing papers has been fully explained to the students; otherwise it would amount to an irrelevant activity without any practical or conceptual application (Dawson 2023; see also the contributions to Dasgupta, Llinas, Gillespie, Fouse, Lawless, Mittu, and Sofge, ed., 2024; these papers discuss the integration of AI into research teams and the high necessity of observing trust and integrity in the process).

Aleksandra Sushchenko and Olena Yatsenko now point out the considerable danger of AI for Higher Education (if not for Education at large), reminding us that already Friedrich Nietzsche (1844–1900) had warned his readers about the threat to human creativity, individuality, and freedom resulting from technology that could overtake human society and enslave it (cf. also Branston 2023). With tongue in cheek, we could admit that AI is obviously a fantastic tool that can make our lives so much easier and smoother, allowing us to leave most of our worries simply behind. Journalists would have to enter only some key words and then would publish whatever article the computer would have produced, without any fact-checking, for instance. As beautiful and clean a paper produced by AI can look like, it eliminates the human dimension, the spontaneity, the creativity, the unexpected, and the spiritual character, replicating what other authors have already formulated without reaching any new insights, putting us really into a vicious cycle without moving beyond in our intellectual pursuit. As Sushchenko and Yatsenko conclude:

this drive toward perfection risks diminishing the richness of ethical life, which depends on the ability to acknowledge moral ambiguities, make mistakes, and learn from them. Nietzsche warns against the suppression of individual will and creativity, arguing that true ethical engagement arises from the freedom to err and reflect on those errors. Therefore, the ethics of technology should not aim to eradicate human flaws but instead recognize the irreplaceable role they play in shaping moral consciousness and fostering a deeper understanding of what it means to be human. This perspective calls for a technological future that preserves room for human spontaneity, allowing for mistakes that lead to growth, rather than imposing rigid moral codes that reduce human experience to mechanical precision.

This is not to say that human errors, failures, misunderstanding, miscommunication, deceptions, lies, or falsifications are to be glorified; on the contrary. However, all those problems indirectly refer to the human traits that make all statements or products so valuable and unique especially in a learning situation; they are not machine-made (cf. Lv, Huang, and Huang, 2023); instead, they are the results of our thinking, our personality and character, or rather the result of our faults, mistakes, ignorance, and confusion. Speaking and writing make us human, allow us to be people as the divinely created beings here on earth, to formulate it in a religious manner of speech. According to Sushchenko and Yatsenko, working through Nietzsche’s concept of humanity and the power differentials present in virtually all societies,

allows us to explore how resistance—existing in creative culture—can serve as a vital counterbalance to the mechanization of social life. Such dialectics provide a strong foundation for supporting algorithmic resistance by inspiring ethical frameworks rooted in individuality and emotional depth, challenging the homogenizing tendencies of digitization and algorithmization. It emphasizes the importance of subjective stories, emotions, and compassion, forming human-centered ethical principles that preserve the richness of individual experiences and protect against system-driven delusions (Abstract).

The future might turn this around, and the robot might become our friend, intimately corresponding with us as needed, and then even reliable because it might reach the level of independent thinking and even feelings. However, at the current stage, turning to AI as the critical tool to produce a text that can be submitted for inspection, grading, or for reaching an agreement (contract), constitutes a considerable danger that undermines the entire purpose of education in the first place. The use of AI also threatens our ethical standards because it invites us to claim authorship when there is none.

Our students learn almost as much from their own mistakes as from instructions because our errors or failures serve like mirrors we can hold up to ourselves and use for introspection, self-criticism, and then improvement, i.e., learning. Relying on a computer to do the job we are required to do and we want to be proud of basically destroys all the joy and pride we gain when we do our own homework.

A simple illustration would be the genre of letters. Writing letters used to be a commonly practiced art, and many of the most famous intellectuals or artists throughout time are best known to us through their correspondence. Today, most people have turned to emails, but those electronic messages rarely serve effectively to relay one’s deeper feelings of love, fear, hope, or aggression. It is imaginable that an individual embarks on a love relationship brought about by AI-generated letters, but it seems highly questionable whether the beloved partner will ever believe the artificially created statements about the other person’s true sentiments. The point here is that AI does not create authenticity; in fact, it robs us of our own self, our self-reliance, and our intellectual ability to produce reasonable and rational texts.

AI and Medieval Literature

A Case Study

Let us consider some concrete examples from an ordinary university class setting. In one of my courses, we were discussing Gottfried von Strassburg’s courtly romance, Tristan and Isolde (ca. 1210; for an excellent English translation, see Gottfried von Strassburg, 2020). This romance stands out for many reasons, such as the hopeless love between the King of Parmenie, Tristan, and the Irish Princess, Isolde, who is married to the King of Cornwall, Mark. This love is triggered, at least superficially, by a love potion which her mother had concocted for her daughter in preparation for her marriage with that older man. We have, of course, to read this potion as a metaphor to explain the sudden change of feelings in Isolde from hating Tristan who had killed her uncle Morold to loving him as the most ideal character in her life. Inadvertently, however, the two young people drink that potion, which then immediately demonstrates its effect. The rest of the romance traces their efforts to be together, to hide their love from King Mark and the court spies, to pretend in public that they are falsely accused, and yet to live out their intimate happiness in secret.

Eventually, they are expelled from the court because Mark no longer believes them although he lacks conclusive evidence of their adultery. Tristan knows of a mysterious cave which ancient giants had built and which allows true lovers like them to exist there in the wilderness far removed from human society without any concerns for their physical well-being. Indeed, the two people enjoy themselves freely and fully, until one day they are discovered by the king, who is, however, deceived once again and made to believe in their innocence because Tristan had placed a naked sword between himself and Isolde in their shared bed. Mark then allows them to return to the court but begs them to watch themselves and to avoid any suspicious gestures or gazes. He wants to believe that he can trust his wife, but at the end, he discovers them in flagrante, which forces Tristan to depart from the court and to roam the world, while he must leave grieving Isolde behind. Gottfried, as the author and narrator, then has Tristan encounter another maid also called Isolde, who hopes for Tristan’s love, but the romance concludes as a fragment, and we as readers are asked to think about the outcome and the many implications by ourselves.

This romance is determined by the experience of happy love, tragic love, desperation, jealousy, fear, and hatred because of the older man being duped, or rejected by his own wife (for an excellent critical discussion, see Tomasek, 2007; cf. also Chinca, 1997; for a variety of critical approaches, see Hasty, ed., 2003). Gottfried von Strassburg was neither the first nor the last to write a story about these two lovers, and the annals of literary history are filled with references to subsequent versions in many different languages, not even considering the many different art works drawing inspiration from this and other texts (cf. the contributions to Eming, Rasmussen, and Starkey, ed., 2012; cf. also my listing of most of the relevant texts online at: https://sites.arizona.edu/aclassen/the-tristan-tree/; last accessed on Jan. 17, 2025). This romance has proven to be a timeless medieval ‘classic’ because it continuously appealed to countless audiences until the sixteenth century, and then again since the late eighteenth century.

Today, whenever the focus rests on issues of love, sexuality, and marriage, and this both in a medieval and modern context, Gottfried von Strassburg’s romance offers itself almost naturally for inclusion in the reading list for a college course. In my case, the course deals with “Eroticism and Love in the Middle Ages” and includes many other medieval texts (Apollonius of Tyre, Andreas Capellanus, troubadour and troubairitz poetry, Marie de France, Carmina Burana, fabliaux and mæren, Oswald von Wolkenstein, and Christine de Pizan; cf. Classen 2025). One of the major tasks for students is to write one, two, or three critical papers with different cues. In this case, though it might sound a bit vague, the syllabus contained this statement:

Create a title, then develop a thesis statement (ca. 50 words), then offer the argument, (ca. 150 words), conclude (ca. 50 words), and have a bibliography: 3 monographs and 3 scholarly articles from the last 10 years at most. One of the articles in a scholarly journal you must have read and integrated into your argument (supporting your thesis or not; you are also invited to consult any of my articles if pertinent). You can choose either: Gottfried’s Tristan, Dietrich of the Glezze (Gletze), or Heinrich Kaufringer.  If you find it valuable, please use “The Knight” as your main text. But in that case, the bibliography might be more difficult. Search under the category of “maeren.” Grading (always up to a max of pts): Thesis: 10 pts. Argument: 50 pts. Conclusion: 10 pts. Bibliography: 15 pts. Inclusion of one article: 15 pts. (https://sites.arizona.edu/aclassen/class-schedule/).

Students had thus much freedom to develop their own thesis to argue about a specific aspect in this romance (or in two or three other shorter texts). However, the requirements also entailed a bit of research, at least in mechanical terms. Unfortunately, ca. 12 out of 85 students submitted papers that were clearly created by AI. Both my Teaching Assistants and I recognized this problem without major difficulties and gave a 0 as a grade. None of those students contested this decision and hence admitted silently that they had indeed used AI for their task. We invited them to submit a truly authentic paper to gain at least partial points. For the purpose of this paper, let’s look at some of the characteristic statements that revealed immediately that a machine had produced them rather meaninglessly, and this in the sense that there was much verbiage that sounded profound and insightful but was nothing like that.

Student Papers Based on AI

Student A begins the paper with this comment: “Gottfried’s Tristan presents a nuanced exploration of love and fate, challenging the medieval notions of chivalric duty and moral integrity. The text reveals how the all-consuming passion between Tristan and Isolde defies societal expectations, suggesting that true love exists beyond human laws and conventions.” Student B used a somewhat different wording, but in essence, it was the very same, again without saying anything specifically drawn from the original text: “In Tristan, Gottfried von Strassburg reveals that love, though deeply desired, is never without hardship. The story explores how the pursuit of love often leads to betrayal, pain, and loss, showing that while love is glorified, it comes with challenges that test the heart and redefine the meaning of devotion.” Student C offers the following introductory statement: “Gottfried von Strassburg’s “Tristan” reinterprets the way we view love by presenting it as both beautiful, but yet love could also be a dangerous and harmful force. Through the characters’ struggles, the text explores the pain, and tension between personal desires, feelings and the rules of society in that time, showing a complex view on the nature of what love really is.”

Student D produced this comment: “The journey of the titular knight in Gottfried’s Tristan shows the conflict between loyalty and personal desire and illustrates how complex chivalric values were. The story highlights the tension between love and loyalty, both essential to the moral code of knights, by looking at his relationships with Isolde and King Mark.” Finally, student E remarks: “In Tristan, Gottfried von Strassburg depicts love as a force that brings upon betrayal, pain, and loss that challenges the idea of love. Through the love potion he explores the idea of love being destructive and beautiful at the same time, which highlights the conflict within a person’s desires, inner self and vies [relies?] on social constructs. The tragedy of Tristan and Isolde demonstrates the painful reality of love and how it clashes with the romanticized version of love. The romanticized idea of love held to high standards by courtly love in the Middle Ages ultimately shows how love is unattainable without consequence and suffering. Tristan and Isolde’s story show the consequences of love through miscommunications, betrayal, and suffering, that truly demonstrates the difference between fanciful love and the painful reality of love. Even well having deep emotional connections [sentence fragment].” Then the student continues: “Their love cannot withstand the difficult situation and will not last in cost [sic] unrepairable loss. Gottfried von Strassburg uses Tristan to portray this idea of doomed love as opposition to the romanticized idea of love. The romanticization of love often causes people to look past the social and emotional complexity of being in a relationship, Overall, the idea of love being both beauty and destruction.”

On the surface, all these summaries seem to make some sense, but at a closer analysis, they amount to the same, virtually meaningless interpretation which is not what the narrative claims at all, especially when we wonder where the topic of social conflicts between the lovers might come from. For instance, arguing that this romance “truly demonstrates the difference between fanciful love and the painful reality of love” amounts to an empty comment because nowhere in the entire romance are there tensions between different types of love, superficial and deep. A sentence like “Gottfried von Strassburg uses Tristan to portray this idea of doomed love as opposition to the romanticized idea of love” just does not make any sense, especially not within the context of courtly love and this famous romance.

The fact that none of these students contested the charge can be taken as admission of guilt (but some specifically confessed it to me). Ironically, these papers offered some references, but at closer examination, it turned out that those mostly do not exist. Thus, from that initial realization, the entire paper then dismantled, leaving the respective student with no defense. In response, we determined that that paper did not deserve any points. However, we encouraged all students who were caught to offer a rewrite based on their own work, with some penalty points. Amazingly, one student submitted a new paper which then turned out to be AI-generated as well.

Old and New Challenges

Altogether, the problem we face here is not new at all and instructors/professors are called upon to embrace, study, or analyze AI themselves in order to handle the imminent challenges (for an excellent discussion of the various approaches we could possibly pursue, see, e.g., this blogsite with practical recommendations: https://blog.cengage.com/part-two-digital-nexus-in-higher-education/?j=2137856&sfmc_sub=911790777&l= 6504_HTML&u=45572007&mid=515011284&jb=1908&utm_ source=marketing&utm_medium=email&mid= 515011284&utm_campaign=fa24_awrn_ nurture&utm _ content=3505684&segment= Faculty®ion=HED_US). Students throughout times have always tried to take the easy route and to comply with the requirements as fast as possible without much effort on their own part. The result was, however, that they lost much more than they had hoped to gain, which might have become much worse today in the current educational situation especially because AI promises to produce a meaningful text that cannot be traced to any source.

The problem with AI-generated papers will continue to grow, especially because the technology will improve. So, what would be some of the solutions that I could imagine or suggest?

Instructors must include a very clear statement as to their policy regarding AI-generated texts.

Instructors must make great efforts to give clearest instructions about the paper and what it is expected to demonstrate and present.

Instructors ought to discuss the very purpose of a college-level paper. The students who relied on AI simply hurt themselves by getting out of the requirement as fast as possible and with the highest possible grade. The more they tried that, of course, the more points they lost. A paper is a very different matter than an exam, a quiz, or an oral report or presentation. To write a paper entails the development of a thesis that then is confirmed by solid arguments, and both of which then allow the author to reach a convincing conclusion. In my case, students were also expected to investigate some of the research available easily, but they did not yet have to engage deeply with it. When a student has to produce a paper, s/he is asked to think for him/herself about the meaning and relevance of a specific text, artwork, political issue or economic problem. The outcome of that paper should mirror thought processes and demonstrate the student’s growing ability to formulate his/her own ideas and to convey a convincing message. By resorting to AI, all that valuable effort was circumvented and avoided.

Conclusion

But I would not want to conclude this paper without offering additional, constructive perspectives. There are many ordinary situations in which an instructor needs to formulate concepts, rules, questions, or other statements that simply reiterate previous comments and only invite the reader to respond according to what s/he has learned. AI might offer an easy and unproblematic solution in this case. A true paper in an academic sense, by contrast, is expected to develop novel ideas, new arguments, and to explore innovative perspectives. I do not doubt that AI will increasingly become useful and helpful in the first case. I strongly question, however, the role of AI in personal growth as a critical thinker and writer because it substitutes the individual rationality and learning process with a mechanized operation that only replicates what can be found in many different manifestations online but does not reach beyond them.

True and independent papers produce authenticity, innovation, and contribute to the creation of inventions and paradigm shifts. AI-generated papers, at least at the present moment, still only convey the pretense that a smart argument was formulated that could pass master. Unfortunately for those who attempt to ‘cheat’ that way, any close reading of such texts quickly reveals this deceptiveness and illustrates how much the alleged writer has become a victim of self-illusion (cf. Collins 2017; see also the wide range of contributions to Czerniewicz and Cronin, ed., 2023; Weng 2024).

Realistically speaking, however, AI as we know it currently is going to develop very fast and in leaps and bounds and might achieve a level of maturity we can only dream about today. Many companies, logistic firms, organizations, airlines, or train systems have already learned the great advantages of AI because it is employed in areas where mechanization and robotization are of great value and use. But higher education does not necessarily aim for faster or better output because we are dealing with young people who are still in the process of learning to express themselves and need exploratory situations that allow them to fathom the range of words or syntactical structures available at large or that would be necessary for their own ability to formulate their own ideas. As with all technologies, there are advantages and disadvantages, and the better we understand both sides, the better we can interact with AI and use it responsibly also in the future. If you cannot fight the enemy, go with him, i.e., the new invention.

To be sure, particularly young learners are not served well with AI because that operational system substitutes for them the actual learning process and impoverishes their intellectual abilities. In classes where we educate our students to become critical and independent thinkers who know how to express their ideas, values, and concepts both orally and in writing, AI primarily serves to transform our human culture into a robotized and mechanized world where we are no longer master of our own destiny  – if we ever could have claimed that fully – but basically just the proverbial cogs in the wheel.

The few examples discussed above have shown clearly how the availability of virtually infinite numbers of websites to AI makes possible for the machine to compile for the users at random or upon cues any kind of text that sounds like a reasonable argument but which, at a closer analysis, mostly turns out to be nothing but rhetorics of a sophist through which no innovative or creative idea is engendered or expressed. This is not to say that we should, or even could reject AI, especially in many areas of human activities or studies that require a lot of mechanical operations. There are many research areas where big data mining is of central concern, such as in the case of linguistic analysis. But there are very concrete limits, and critical thinking and academic writing are not to be replaced by artificially produced texts, especially when the individual student is supposed to learn how to respond to a text, an image, a psychological situation, an economic conflict, or a political dispute, and this in writing.

Of course, one could certainly use AI-generated texts for comparative purposes to illuminate the advantages and disadvantages, but in practical terms, most students, especially on the undergraduate level, are best advised and strongly urged not to utilize AI because they do not know how to handle it responsibly. When we realize, as it is happening right now, the dangerous tendency of AI to create fake news, fake statements, fake comments, and faulty interpretations, then AI turns out to be a dangerous tool for students since they become subject to a technological system that is not at all reliable, that is manipulable, and openly misleading, often for political, religious, or generally ideological purposes. Relying on AI means relying on what has been formulated so far, mixing and restating it by means of a robotic mechanism. AI stifles all human creativity, at least in the Humanities and Fine Arts, but certainly also in many neighboring disciplines.

But AI is here to stay with us, and we cannot assume that in the future students will not draw from it because of the ease and speed with which AI produces papers. Hence, we need to challenge our students in new ways, confronting them with samples of traditional writings compared with those produced by AI, allowing them thus to learn how to distinguish between human- and machine-based narratives. We also need to rethink the format and nature of the assignments for our students. The ultimate question hence will be how we will be able to manage the beast which we ourselves have created and put it to the best possible use where imaginative, creative, intelligent forces are not required. Let AI crunch big data, let it be a servant for dull, mechanical tasks, let it help us to correct grammatical and spelling issues (as in the case of my own article), but we must be on the lookout when or if it replaces human intelligence, creativity, and critical thinking. Education aims at the individual, and the individual must gain independence, and this based on personal knowledge and critical thinking skills. No AI can or should ever replace this fundamental experience transforming a young person into a mature adult.

It might well be that AI can assist younger students well as machine tutors, as Hartley, Hayak, and Ko (2024) now argue. But it depends greatly on the specific circumstances, the learner’s motivation, and the concrete tasks pursued by this tutoring process. As they themselves admit, with which I want to conclude this paper: “Individualization is a hallmark of human tutoring and has implications for each area explored. The anticipation is that this can be a strength of AI systems. However, at this juncture, it is a challenge. While ChatGPT can follow the thread of the conversation, it is not consistent or robust. This is likely due to the probabilistic nature of the existing system’s dependence upon large language models and predictive algorithms. This approach emphasizes the probability of a plausible response over the likelihood of a helpful answer.”

Indeed, the human dimensions in writing, in critical assessments of narratives, of artwork, or philosophical treatises continue to be of central importance for the learning process at every stage in life. I am even afraid that AI might threaten or undermine the individual interest, inspiration, and motivation so essential for students who must learn to become independent thinkers, writers, or artists. However, the flipside might be that a smart employment of AI could achieve the very opposite goal, much depending on our human input and coordination with the machine within education. Nevertheless, learning ultimately always takes place within the individual, irrespective of the technical tools that might assist in that process. We teach because we want to educate the new generation to become independent thinkers, logical and rational writers, and ethical and moral individuals, and this by way of engaging, at least in my case, with major literary works from the past and present.

Acknowledgement

I wish to express my gratitude to the University of Arizona where I research and teach for making all the research and library tools necessary for this project available to me.

Funding Sources

The author(s) received no financial support for the research, authorship, and/or publication of this article

Conflict of Interest

The author declares no conflict of interest.

Data Availability Statement

This statement does not apply to this article.

Ethics Statement

This research did not involve human participants, animal subjects, or any material that requires ethical approval, and therefore, informed consent was not required.

Author Contributions

The sole author was responsible for the conceptualization, methodology, data collection, analysis, writing, and final approval of the manuscript.

References

1. Aaron, Lynn, Santina Abbate, and Nicola Marae Allain (2024). Optimizing AI in Higher Education: SUNY FACT Guide. Albany, NY: SUNY Press.
   CrossRef
2. Abelard, Peter (2008). Yes and no: The Complete English Translation of Peter Abelard’s Sic et Non, trans. Priscilla Throop. Charlotte, VT: MedievalMS.
3. Bayer, Judit and Christian Grimme, ed. (2025). Code and Conscience: Exploring Technology, Human Rights, and Ethics in Multidisciplinary AI Education. Cham: Springer Nature Switzerland.
   CrossRef
4. Bezner, Frank (2005). Vela Veritatis: Hermeneutik, Wissen und Sprache in der Intellectual History des 12. Jahrhunderts. Brill Book Archive, 1. Leiden and Boston: Brill.
   CrossRef
5. Binini, Irene (2022). Possibility and Necessity in the Time of Peter Abelard. Investigating Medieval Philosophy, 15. Leiden and Boston: Brill.
   CrossRef
6. Bowen, José Antonio and C. Edward Watson (2024). Teaching with AI: A Practical Guide to a New Era of Human Learning. Baltimore, MD: Johns Hopkins University Press.
   CrossRef
7. Branston, Tyler (2023). AGI, All Too Human; Nietzsche and Artificial General Intelligence. Ph.D. dissertation, University of Victoria, Victoria, BC, Canada.
8. Chan, Cecilia Ka Yuk (2024). Generative AI in Higher Education: The ChatGPT Effect. New York and London: Taylor and Francis.
   CrossRef
9. Chinca, Mark (1997). Gottfried von Strassburg: Tristan. Landmarks of World Literature. Cambridge: Cambridge University Press.
   CrossRef
10. Classen, Albrecht (2025). “The Discourse of Courtly Love in Medieval Verse Narratives,” Encyclopedia 2024, 4, 1904–1917.
    https://doi.org/10.3390/encyclopedia4040124Encyclopedia 4 (2024): 1904–1917; online at: https://doi.org/10.3390/encyclopedia4040124(MDPI)
    CrossRef
11. Classen, Albrecht (2020). “Teaching Medieval Studies with a Modern Learning Management System: Top Hat in a Medieval Seminar,” Once and Future Classroom, XVI (1); https://once-and-future-classroom.org/; or: https://once-and-future-classroom.org/teaching-medieval-studies-with-a-modern-learning-management-system-top-hat-in-a-medieval-seminar-2/.
12. Collins, Harry (2018). Artifictional Intelligence: Against Humanity’s Surrender to Computers. Cambridge: Polity.
13. Czerniewicz, Laura and Catherine Cronin, ed. (2023). Higher Education for Good: Teaching and Learning Futures. Cambridge: Open Book Publishers.
    CrossRef
14. Dawson, April G. (2023). Artificial Intelligence and Academic Integrity. Frederick: Aspen Publishing.
15. Dasgupta, Prithviraj, James Llinas, Tony Gillespie, Scott Fouse, William Lawless, Ranjeev Mittu, and Donald Sofge, ed. (2024). Putting AI in the Critical Loop: Assured Trust and Autonomy in Human-Machine Teams. San Diego: Elsevier Science & Technology.
16. Eming, Jutta, Ann Marie Rasmussen, and Kathryn Starkey, ed. (2012). Visuality and Materiality in the Story of Tristan and Isolde. Notre Dame, IN: University of Notre Dame Press.
17. Gottfried von Strassburg. (2020). Tristan and Isolde with Ulrich von Türheim’s Continuation. Ed. and trans., with an intro. by William T. Whobrey: Hackett Publishing.
18. Hartley, Kendall, Merav Hayak, and Un Hyeok Ko (2024). “Artificial Intelligence Supporting Independent Student Learning: An Evaluative Case Study of ChatGPT and Learning to Code.” Educ. Sci., 14(2), 120; https://doi.org/10.3390/educsci14020120.
    CrossRef
19. Hasty, Will, ed. (2003). A Companion to Gottfried von Strassburg’s “Tristan”. Studies in German Literature, Linguistics, and Culture. Rochester, NY: Boydell & Brewer.
20. Kersten, Saskia and Christian Ludwig, ed. (2024). Born-Digital Texts in the English Language Classroom. New Perspectives on Language and Education, 39. Bristol, UK: Multilingual Matters.
    CrossRef
21. Lv, Linxiang, Minxue Huang, and Ruyao Huang (2023). “Anthropomorphize Service Robots: The Role of Human Nature Traits.” The Service Industries Journal, 43 (3–4), 213–37; online at: https://www.tandfonline.com/doi/epdf/10.1080/02642069.2022.2048821?needAccess=true.
    CrossRef
22. Marenbon, John (1997). The Philosophy of Peter Abelard. Cambridge: Cambridge University Press.
    CrossRef
23. Martin, Meredith (2025). Poetry’s Data: Digital Humanities and the History of Prosody. Chicago and London: University of Chicago Press.
    CrossRef
24. Murugan, Thangavel, Karthikeyan Periasamy and A. M. Abirami, ed. (2025). Adopting Artificial Intelligence Tools in Higher Education: Teaching and Learning. Abingdon, Oxon, and New York, NY: Routledge.
    CrossRef
25. Pan, Fan (2024). AI in Language Teaching, Larning, and Assessment. Hershey, PA: IGI Global.
    CrossRef
26. Sushchenko, Aleksandra and Olena Yatsenko (2025). “Human, All Too Human: Do We Lose Free Spirit in the Digital Age?” Humanities, 14 (6), online at: https://doi.org/10.3390/h14010006
    CrossRef
27. Tomasek, T. (2007). Gottfried von Straßburg. Stuttgart: Philipp Reclam jun.
28. van Lit, L.W. C. (2019). Among Digitized Manuscripts. Philology, Codicology, Paleography in a Digital World. Leiden and Boston: Brill.
    CrossRef
29. van Renswoude, Irene (2024). “After Digitization, What More?: The Touch and Feel of Virtual Manuscripts.” Quaerendo, 54(2–3), 210–25.
30. Vanderheiden, Elisabeth and Claude-Hélène Mayer, ed. (2020). Mistakes, Errors and Failures Across Cultures: Navigating Potentials. Cham: Springer.
    CrossRef
31. Weng, Wei (2024). A Beginner’s Guide to Informatics and Artificial Intelligence: A Short Course Towards Practical Problem Solving. Singapore: Springer Nature Singapore.
    CrossRef
32. Yu, Jiyuan and Jorge J. E. Gracia, ed. (2003). Rationality and Happiness: From the Ancients to the Early Medievals. Rochester Studies in Philosophy, 4. Rochester, NY: University of Rochester Press.
33. Yuan, Ye, Huan Li, and Anan Sawaengdist (2024). “The Impact of ChatGPT on Learners in English Academic Writing: Opportunities and Challenges in Education.” Language Learning in Higher Education, 14(1), 41–56.
    CrossRef
34. Error-prone AI-made news summaries suspended in beta iPhone software. (Jan. 18, 2025). Retrieved Feb. 11, 2025, from https://rapidcityjournal.com/error-prone-ai-made-news-summaries-suspended-in-beta-iphone-software/article_c89e36c2-355c-54fb-8549-f51afea89d13.html
35. Meta to end fact-checking, replacing it with community-driven system akin to Elon Musk’s X. (Jan. 8, 2025). Retrieved Feb. 11, 2025, from https://www.cbsnews.com/news/meta-facebook-instagram-fact-checking-mark-zuckerberg/
36. What’s the impact of artificial intelligence on energy demand? (Nov. 17, 2024). Retrieved Feb. 11, 2025, from https://www.sustainabilitybynumbers.com/p/ai-energy-demand
37. Artificial intelligence’s thirst for electricity. (July 10, 2024). Retrieved Feb. 11, 2025, from https://www.npr.org/2024/07/10/nx-s1-5028558/artificial-intelligences-thirst-for-electricity