Returning to Socrates: Why Generative Artificial Intelligence Compels Us to Reinvent the University

The emergence of generative artificial intelligence is forcing universities to reconsider some of the most fundamental questions about teaching and learning. If a machine can write an essay, summarize a book, or provide a sophisticated answer to a complex question in a matter of seconds, what is the point of continuing to teach and learn in the same way as before? This article argues that the real challenge is not technological but pedagogical. The issue is not whether universities should prohibit or embrace artificial intelligence, but rather how they can rediscover what has always been their essential mission: to educate individuals capable of understanding, reasoning, deliberating, and exercising independent judgment.

Paradoxically, responding to this challenge does not require inventing an entirely new pedagogy. Instead, it calls for recovering some of the richest intellectual traditions of the Western university. From the Socratic dialogue to the lectio, quaestio, and disputatio of the medieval university, higher education was built for centuries upon the art of questioning, debating, and responding to objections. At a time fascinated by the speed with which answers can be produced, perhaps the most profound educational innovation is to teach once again what no artificial intelligence can do on our behalf: to think autonomously, engage in critical dialogue, and transform information into genuine judgment.

Introduction

Only a few years ago, assigning students an essay or research paper was a reasonably reliable way of determining whether they had understood a subject. Writing required them to read extensively, select appropriate bibliography, organize ideas, discard weak arguments, and construct a line of reasoning of their own. The final paper might be better or worse, but it almost always reflected an intellectual process that instructors could reasonably assume had taken place. Today, that assumption is no longer self-evident.

A scene is beginning to repeat itself in universities across the world. A professor assigns, for example, an essay on John Rawls's theory of justice. One week later, thirty impeccably written papers are submitted. They are well structured, supported by appropriate references, and written in remarkably uniform prose. Nothing initially suggests that there is any problem. Yet as soon as the professor begins asking students about the ideas they have just presented, it becomes apparent that several of them are unable to explain why they defended a particular interpretation, why they cited certain authors rather than others, or what objections might be raised against the central thesis of their essays. The written work appears excellent, but the reasoning that ought to sustain it proves surprisingly fragile.

The most immediate reaction is to interpret this situation as a problem of academic misconduct. Discussions quickly focus on artificial intelligence detectors, new disciplinary regulations, or systems capable of determining whether a text has been written by a human being or generated by an algorithm. Such measures may well be necessary, but they are clearly insufficient. The real challenge is not simply to determine who wrote a paper. Rather, it is to recognize that the relationship between writing and thinking has fundamentally changed and to draw the pedagogical consequences of that transformation. It has become remarkably easy to produce an academically acceptable text without necessarily undergoing the intellectual journey that such writing once required.

This is the great paradox of generative artificial intelligence. Systems such as ChatGPT, Claude, and Gemini can summarize scholarly articles, write essays, compare philosophical theories, and suggest relevant bibliography in a matter of seconds. They are extraordinarily valuable tools for studying, conducting research, and exploring new ideas. The problem lies neither in their existence nor in their legitimate use, but in the temptation to confuse the quality of an answer with the quality of the thinking that ought to have produced it. Artificial intelligence has not rendered writing obsolete; rather, it has broken the long-standing equivalence between writing and thinking.

The more sophisticated machines become at producing answers, the clearer it becomes that the fundamental mission of the university has never been merely to transmit information. Since its origins, the university has pursued a far more ambitious goal: to educate individuals capable of understanding, reasoning, deliberating, and exercising independent judgment. Artificial intelligence does not undermine this mission; on the contrary, it makes it more visible than ever.

Perhaps, then, we have been asking the wrong question. Instead of asking only how universities should adapt to artificial intelligence, we should first ask what remains worth teaching when a machine can answer, in seconds, questions that once required hours of study. At precisely this point, an old Athenian philosopher returns to the conversation - a man who never wrote a book and devoted his entire life to asking questions. The future of the university may depend less on learning how to use artificial intelligence than on recovering what Socrates regarded as the very essence of education: teaching people how to think.

Every Technological Revolution Compels Us to Reinvent the University

The feeling that the university is undergoing an unprecedented transformation is entirely understandable. The extraordinary pace at which artificial intelligence has evolved gives the impression that we are witnessing a radical break with everything that has come before. Yet the history of education invites us to adopt a broader perspective. Universities have endured for centuries precisely because they have never identified their mission with any particular technology. Their tools have changed, as have their ways of organizing knowledge, but one enduring question has remained the same: what is truly worth learning?

The invention of the printing press gave rise to fears that now seem almost quaint. If books could be reproduced on a massive scale, would it still make sense to memorize so many texts? The emergence of great libraries profoundly transformed humanity's relationship with knowledge, even though no student could hope to master everything that had been written. More recently, the Internet made access to information largely independent of physical proximity to a university library. Each of these developments prompted predictions about the decline of traditional education. Yet what actually happened was quite different. Universities came to understand that whenever access to knowledge changes, teaching and assessment must also evolve.

Artificial intelligence represents another chapter in this long history, but with one crucial difference. The printing press multiplied books; the Internet multiplied information; artificial intelligence multiplies answers - and even entire texts. That is the true novelty. It does not merely facilitate access to existing knowledge. It produces texts, organizes arguments, summarizes research, translates articles, and generates explanations with a speed that would have seemed unimaginable only a few years ago. For the first time, a machine is not simply storing information; it is participating in activities that have traditionally been regarded as belonging to intellectual work itself.

Faced with this new reality, it is understandable that many universities have responded by establishing regulations governing the use of these tools or by developing systems to detect AI-generated texts. Such measures are undoubtedly necessary, but they do not address the heart of the problem. The fundamental challenge is neither disciplinary nor technological; it is pedagogical. Every technological revolution forces us to reconsider which human capacities remain uniquely our own and, consequently, what the university should continue to cultivate. Artificial intelligence confronts us with an even more radical question: what does it mean to learn when almost any answer can be obtained within seconds?

The easier it becomes to access information, the less important it is merely to accumulate it, and the more essential it becomes to understand it. An abundance of answers does not diminish the need for judgment; it makes that need more urgent than ever. Knowledge is no longer simply a matter of finding information, but of interpreting it, connecting it, evaluating it critically, and deciding which sources deserve our trust. The university has never been merely a distributor of information. Its enduring mission has always been to cultivate minds capable of finding their way through complexity. It is precisely here that the debate about artificial intelligence ceases to be primarily technological and reveals itself as a profoundly philosophical question.

For if the real challenge is to teach people how to think rather than simply how to answer, there may be no need to invent an entirely new pedagogy. Instead, we may need to return to the thinker who made questioning, dialogue, and the critical examination of one's own beliefs the very foundation of education. In this sense, artificial intelligence compels us to return to Socrates.

The Philosopher Who Taught People to Think Without Giving Them Answers

When Socrates is mentioned, the image that usually comes to mind is that of an elderly man walking through the streets of Athens, engaged in conversation with his companions. Yet Socrates did not teach in the sense in which we understand teaching today. He did not lecture, dictate notes, or present a body of knowledge for his students to memorize. Instead, he did something far more unsettling: he asked questions. He questioned his interlocutors until answers that had initially seemed obvious no longer appeared so. He questioned them until they discovered contradictions they had never noticed before. In short, he questioned them in order to compel them to think.

This method bewildered many of his contemporaries because it reversed the traditional role of the teacher. The teacher was no longer the person who possessed all the answers, but the one who revealed the inadequacy of easy answers. Socratic irony did not consist in ridiculing one's interlocutor or demolishing their arguments. Rather, it exposed the fact that genuine understanding demands far more than simply repeating a belief or recalling a piece of information. Learning begins precisely when we realize that we do not yet understand what we thought we knew.

For centuries this approach to education may have appeared slow, even inefficient. Why devote so much time to conversation when one could simply provide the correct answer? Modern education has, to some extent, inherited this impatience. It has often equated learning with the accumulation of information and teaching with transmitting that information as efficiently as possible. Yet artificial intelligence is now exposing the limitations of this conception. If a machine can provide an accurate explanation of almost any subject within seconds, then the professor's role can no longer be reduced to delivering information that students could easily obtain elsewhere.

It is precisely here that Socrates regains his extraordinary relevance. What no artificial intelligence can do for a student is undertake the journey toward understanding on that student's behalf. AI can summarize Kant's Critique of Pure Reason, explain the theory of evolution, compare Aristotle's and Rawls's conceptions of justice, or even produce a polished essay on any of these subjects. What it cannot do is replace the moment when an individual discovers for themselves why one argument is more persuasive than another, recognizes the weaknesses in their own reasoning, or revises a deeply held conviction after encountering a more compelling objection. That intellectual transformation remains irreducibly human.

It is therefore worth remembering that Socrates never sought to fill his disciples' minds with ready-made answers. His ambition was far greater: he sought to teach them how to remain intellectually awake. The elenchus and the maieutic method were not simply techniques for extracting hidden knowledge; they were ways of accompanying another person until that person became capable of thinking independently. Education did not consist in replacing ignorance with a collection of certainties, but in cultivating an intellect capable of examining every certainty critically - including its own.

This is perhaps Socrates' most enduring lesson for the age of artificial intelligence. AI has dramatically reduced the cost of obtaining answers. Precisely for that reason, the educational value shifts elsewhere: to the capacity to formulate meaningful questions. Asking a good question requires defining a problem, distinguishing what is essential from what is incidental, identifying the assumptions underlying a claim, and recognizing the limits of one's own knowledge. Answers can increasingly be automated; genuinely important questions still arise only from a mind seeking understanding.

The university of the twenty-first century may therefore need to embrace an unexpected paradox. For decades educators worried that students could too easily find answers on the Internet. Today, answers often appear before the question has even been fully formulated. Yet the ability to ask good questions has never been more important. Questions are not merely instruments for acquiring information; they are the means by which thought itself is directed. And teaching people how to direct their thinking has been, for more than twenty-five centuries, the true vocation of philosophy.

Reinventing the University in the Age of Artificial Intelligence

When discussing artificial intelligence, the future of the university is often presented as a choice between embracing these new technologies or prohibiting them. This is a false dilemma. History shows that educational institutions have never survived by rejecting technological innovation, but by learning how to incorporate it without losing sight of their fundamental purpose. The printing press did not bring teaching to an end, nor did the Internet make professors obsolete. Likewise, artificial intelligence will not eliminate the university. What it can do is compel the university, once again, to ask what its true mission is.

That mission is to teach students to distinguish between information and knowledge, between knowledge and understanding, and between understanding and judgment. Artificial intelligence can assist us in the early stages of this journey, but its final stages still depend upon personal reflection, dialogue with others, and the intellectual responsibility of the individual thinker. Significantly, the pedagogical methods required to cultivate these capacities need not be invented from scratch. Many of them already belong to the university's own tradition.

The intellectual life of the medieval university, for example, revolved around three complementary practices: the lectio, the quaestio, and the disputatio. Learning began with the careful reading of authoritative texts, continued through the formulation of genuine problems, and culminated in reasoned debate, where knowledge was tested through dialogue, argumentation, and responses to objections. In an age in which artificial intelligence provides immediate access to answers, these practices have lost none of their relevance. On the contrary, they offer a particularly valuable model for cultivating critical judgment and intellectual autonomy.

The real danger, therefore, does not lie in using artificial intelligence, but in allowing it to perform those tasks that constitute the very core of our intellectual independence: understanding, reasoning, making decisions, and accepting responsibility for our own judgments. A university that abandoned the task of forming this intellectual sovereignty might continue to award degrees, but it would have forgotten the mission that justified its existence in the first place.

The appropriate response to artificial intelligence is not to compete with machines, but to cultivate precisely those capacities that machines cannot replace: reading attentively, asking questions that go beyond the answers readily available, engaging in genuine conversation, listening seriously to opposing arguments, changing one's mind when the evidence requires it, and learning to write as a way of learning to think rather than merely producing text. These were, to a remarkable extent, the educational intuitions that guided Socrates twenty-five centuries ago, and it is perhaps for that very reason that they have acquired such unexpected relevance today.

Paradoxical though it may seem, the most advanced technology humanity has ever created is compelling us to rediscover some of the oldest educational practices of our intellectual tradition. The more capable machines become of generating answers, the more important it becomes to educate people capable of asking meaningful questions. The easier it is to produce information, the more necessary it becomes to interpret it with discernment. And the more powerful the technologies that extend our cognitive abilities become, the greater our responsibility to preserve what no technology can exercise on our behalf: judgment.

Perhaps this is one of the most valuable contributions that artificial intelligence can make to the contemporary university. It does not require the university to abandon its past or begin again from scratch. Instead, it invites the institution to recover the very traditions that have always defined its highest educational aspirations.

Ernesto Baltar

Senior Lecturer in Philosophy at Rey Juan Carlos University

Follow Ernesto Baltar on LinkedIn