Author: Alejandro Díaz Garreta

  • Artificial Intelligence Is Not the Problem. Education Is.

    A Misplaced Anxiety

    Much of the current debate about artificial intelligence in education begins in the wrong place.

    We worry that students may use AI systems to complete tasks once treated as evidence of learning: essays, problem sets, summaries, and explanations. We ask how to prevent misuse, preserve integrity, and defend established forms of assessment.

    These concerns are real. But they do not reach the deeper issue.

    Artificial intelligence did not create the gap between performance and understanding. It exposed it.

    What appears as technological disruption may reveal something older and more structural: a weakness in how education has often equated visible output with genuine formation.

    The Proxy Assumption

    For long periods of educational history, producing the answer usually required substantial intellectual effort. There was no easy alternative source. Under those conditions, visible performance often functioned as a reasonable proxy for underlying understanding. But a proxy is not an identity.

    To produce a correct result is not necessarily to possess the capacities that the result appears to display. Correct outputs may arise from memorisation, imitation, procedural training, narrow rehearsal, or temporary pattern recognition. They may indicate learning, but they do not guarantee it.

    What artificial intelligence has changed is not the truth of this distinction, but our ability to ignore it.

    Tasks once sustained by the practical difficulty of imitation can now be completed fluently, instantly, and at scale. The proxy assumption, long tolerated, has become unstable.

    Artificial Intelligence as a Diagnostic

    Artificial intelligence functions less as an enemy than as a diagnostic. It reveals what our tasks were demanding of learners.

    When a generative system can complete an assignment with little apparent loss of value, a serious question arises: what, precisely, was the task cultivating in the learner?

    This does not mean the task was worthless. Many traditional exercises built fluency, discipline, familiarity, and background knowledge. But their educational value may not have been the value often attributed to them.

    Activities treated as evidence of thought may have depended more heavily on execution than on judgement.

    AI now performs execution with increasing competence. In doing so, it forces institutions to distinguish what can be produced from what must be formed.

    Result and Formation

    The present moment makes one distinction difficult to avoid: the distinction between obtaining a result and being changed by the effort of reaching it.

    A student may submit a correct mathematical solution generated by AI. Yet when asked why a particular step was necessary, how the reasoning would change in a new case, or where the method would fail, uncertainty appears. The result exists. The underlying capacity may not.

    This does not mean formation occurs only through solitary struggle from first principles.

    Students often learn through examples, models, imitation, and guided solutions. A proof carefully studied can teach more than one poorly invented. An elegant solution, when examined attentively, may cultivate greater understanding than blind trial and error.

    But in such cases, formation occurs only when the learner reopens the process: reconstructing reasons, testing alternatives, identifying limits, and making the logic their own.

    Outputs may support formation. They do not automatically constitute it.

    An Older Warning

    In the Phaedrus, Plato recounts a warning about writing. Writing, he is told, may give the appearance of wisdom without its reality.

    Historically, Plato was too pessimistic. Writing enabled forms of thought impossible without it. Philosophy, science, law, and memory itself were transformed by textual culture.

    Yet the warning still contains an enduring insight. A technology may extend intelligence while also loosening the connection between possession and acquisition, between access to knowledge and earned understanding.

    Artificial intelligence raises a related question today. It can generate forms associated with learning (explanation, argument, solution, summary) without ensuring that the learner has undergone the intellectual work those forms once implied.

    What Education Must Now Clarify

    If intelligent systems can generate answers instantly, education faces a structural choice.

    It can concentrate primarily on the use of the policing tool. Or it can clarify what education was meant to cultivate all along.

    Understanding becomes visible when a learner can reconstruct reasoning in a new situation, recognise where a method fails, and defend conclusions that are genuinely their own.

    These capacities belong to judgement. They may be assisted by tools, but they cannot be substituted by them.

    Conclusion

    Artificial intelligence did not break education. It revealed where education had confused outputs with understanding, performance with formation, and visible success with the formation it was meant to produce.

    The central task now is not to preserve familiar assignments merely because they are inherited.

    It is to design forms of learning in which result and formation cannot be cleanly separated — where producing the answer already requires becoming capable of understanding it.

  • From Turing to Delegated Thinking. How recognition reshaped thinking — and why it matters for education

    A Displacement in the Question of Thinking

    In 1950, Alan Turing proposed what would become one of the most influential moves in the philosophy of mind. Instead of asking what it means to think, he suggested replacing the question with another: under what conditions would we recognise that thinking is taking place?

    This shift appears modest. It does not deny that thinking exists, nor does it attempt to define it. It simply relocates the problem from the nature of thinking to the criteria by which it is identified.

    Yet this relocation is not neutral. It introduces a structural transformation. The question of thinking is no longer tied to an internal process, but to an external judgment. Thinking becomes something not known directly but inferred from observable outputs.

    This is the origin of a displacement whose consequences extend far beyond Turing’s immediate context.

    From Definition to Recognition

    Once thinking is framed in terms of recognition, its criteria become operational. The problem is no longer to understand what thinking is, but to determine when behaviour is sufficient for it to be attributed.

    This is the logic of the Turing Test. If a system produces responses indistinguishable from those of a human, we treat it as if it were thinking.

    The move is powerful because it avoids metaphysical debates. It does not require agreement on the nature of the mind. It only requires agreement on what counts as evidence.

    But in doing so, it changes the structure of the problem. Thinking is no longer grounded in its origin, but in its appearance.

    Recognition and the Formation of Trust

    When outputs are taken as the basis for recognising thinking, a second shift follows.

    If a system consistently produces responses that satisfy the criteria of recognition—coherence, relevance, correctness—it becomes reasonable to rely on those outputs. Not because we have resolved the question of whether the system thinks, but because its performance is sufficient for our purposes.

    Trust emerges at the level of adequacy.

    This trust does not remain passive. Repeated reliance on outputs that prove consistently adequate produces a shift in the structure of activity itself. The system is no longer consulted only in exceptional cases; it becomes integrated into the process through which tasks are carried out.

    At first, it supports reasoning. Gradually, it begins to organise itself.

    From Trust to Delegation

    It is at this point that a more subtle transformation occurs.

    The issue is no longer whether the system thinks, nor whether its outputs are reliable. The issue is how its presence alters the distribution of cognitive activity.

    When a system can produce structured responses across a wide range of domains, the question is no longer simply whether to use it, but when.

    This temporal shift is decisive.

    A tool that is consulted after a problem has been defined supports the execution of thought. A system that is consulted before the problem is fully articulated begins to shape its formation.

    The distinction is not between correct and incorrect answers. It is between those who determine what the question is.

    Why the Point of Intervention Moves

    What distinguishes these systems from prior cognitive tools is not their reliability, but the level at which they operate.

    A calculator intervenes within a problem that has already been defined. Its use presupposes that the agent has determined what is to be calculated. Its reliability does not alter the structure of inquiry because it does not participate in its formation.

    AI systems operate differently. Their outputs do not simply resolve tasks within a given frame; they propose structured interpretations of situations. To consult them is not merely to execute, but to receive a possible articulation of what the situation is a case of.

    For this reason, their point of entry is not fixed. The more reliable these outputs become, the more natural it becomes to consult them precisely at the moment when the situation is still indeterminate. Not because the agent abandons thinking, but because the system offers, at that stage, something that previous tools could not: a candidate form.

    It is this shift—from tools that operate within a frame to systems that propose the frame itself—that allows the point of consultation to move earlier in the process. The change is not only one of frequency, but of position within the structure of inquiry.

    Delegated Thinking

    This transformation gives rise to what may be called delegated thinking.

    Delegated thinking is not the outsourcing of answers. It is the gradual transfer of cognitive initiative from the individual to the system—not through explicit decision, but through habituation.

    At first, the system assists. Then it anticipates. Eventually, it begins to define the terrain within which reasoning takes place.

    The individual continues to evaluate, to select, to refine. But the origin of inquiry—the moment in which a situation becomes a question—shifts elsewhere.

    This is not a loss of activity, but a reconfiguration of it.

    On the Extended Mind

    This account should not be confused with the extended mind thesis, associated with Andy Clark and David Chalmers, which holds that cognitive processes may be distributed across internal and external systems.

    The issue here is not whether cognition can extend beyond the individual, but whether the capacity to initiate inquiry remains actively formed within it.

    A system may participate in reasoning without displacing it. But when it begins to supply not only operations but also candidate forms for what the situation is, its role changes. It no longer extends cognition within a given structure; it intervenes in the formation of that structure itself.

    The question, therefore, is not whether thinking is distributed, but whether the origin of inquiry remains something the individual must bring into being.

    Initiation and the Structure of Inquiry

    To understand what is at stake, it is necessary to distinguish between two dimensions of thinking.

    The first is evaluation: the activity of reasoning within a given frame, assessing alternatives, and arriving at conclusions.

    The second is initiation: the activity through which a situation first becomes a problem for thought.

    Initiation occurs when a difficulty is sensed but not yet articulated. The elements of the situation are present, but their relevance is not yet determined. The mind must impose form on what is still indeterminate.

    This is the moment in which thinking begins.

    Evaluation depends on initiation. One cannot reason about a problem that has not yet been constituted as such.

    Even when the problem’s formation is externally supplied, evaluation can still take place. But it takes place within a space that has already been structured.

    Technological Mediation and the Conditions of Initiation

    Modern AI systems intervene precisely at this level.

    Their outputs do not merely provide answers; they organise possibilities. They suggest what is relevant, how elements are related, and what the situation may be about.

    In doing so, they reduce the indeterminacy that characterises the moment of initiation.

    This reduction is not imposed. It is accepted because it is useful. It accelerates understanding, clarifies options, and often leads to correct conclusions.

    But it also alters the conditions under which thinking begins.

    The situation is no longer encountered as open. It is encountered as already interpreted.

    Education and the Formation of Judgement

    The implications for education are direct.

    If students encounter problems that have already been structured, they may learn to reason effectively within given frameworks. They may produce valid answers, apply correct methods, and demonstrate technical competence.

    But the capacity to determine what is worth asking—to recognise what matters in a situation that has not yet been defined—may remain underdeveloped.

    This capacity cannot be acquired solely through evaluation. It requires exposure to indeterminacy, and the need to resolve it.

    If that moment is systematically pre-empted, the formation of judgement is altered at its origin.

    Conclusion

    The significance of these systems does not lie in whether they think. It lies in how their presence reorganises the structure of human thinking.

    By making recognition sufficient, and by producing outputs that satisfy that recognition, they become reliable partners in cognitive activity. Through repeated use, they shift the point at which they are introduced—from execution to formation.

    What changes is not the presence of thought, but its structure: the relation between what is given and what must be formed.

    It is within this relation that judgement emerges. And it is this relation that is now being transformed.

  • Truth, Opacity, and the Conditions of Thinking

    1. Two Scenes

    In The Brothers Karamazov, Ivan confronts something he cannot resolve. He does not act because he understands, but because he cannot escape what he has seen. The experience is not one of clarity, but of exposure. Something resists him, and in that resistance, something in him is formed.

    In The Trial, Josef K. moves through a world that demands action without explanation. He follows procedures whose meaning is never disclosed. Action persists, but understanding does not. He is not paralysed. He is carried.

    These are not literary examples. They describe two different ways of being placed in relation to what one does not understand.

    2. Two Relations to What Resists Us

    In the first, something interrupts the subject. It cannot be absorbed or bypassed. It must be faced.

    In the second, nothing interrupts. The subject continues to act, but never encounters what exceeds it. The process continues, even if its meaning does not.

    Both involve opacity. But they are not the same.

    In one, something confronts you. In the other, something carries you.

    3. When Nothing Interrupts

    There are situations in which we do not fully understand what we are doing, and yet something still resists us. We hesitate, we return, we revise. The lack of clarity does not remove the encounter.

    But there are others in which this does not happen. The process unfolds. The next step is available. The answer appears. Nothing forces us to stop. Nothing forces us to reconsider where we are, or how we got there. We move forward, but without ever quite encountering what we are moving through.

    4. A Subtle Shift

    This difference is easy to overlook. 

    From the outside, both situations appear similar. In both, the individual does not fully understand. In both, action continues.

    But from the inside, they are not the same.

    In one, the lack of clarity exposes us. In the other, it is absorbed before it reaches us.

    5. Where Thinking Begins

    We often assume that thinking is defined by what we produce: answers, decisions, conclusions. But it may depend just as much on where it begins.

    Whether we encounter something that resists us. Or whether we are carried forward before that encounter takes place.

    6. An Open Question

    If thinking depends on remaining exposed to what we do not yet understand, then a question emerges: What happens when that exposure becomes less frequent?

    Not because the world is simpler, but because we are no longer the ones who encounter its resistance first.

  • Mimetic Desire, Power and the Risk of Delegated Thinking in the Age of AI

    A recent article by Nicolas Truong in Le Monde examines how figures such as Peter Thiel and sectors of the American techno-reactionary movement have appropriated René Girard’s thought. The article is not, strictly speaking, a theory of artificial intelligence or education. It is an exercise in intellectual history. But it contains an intuition whose implications extend far beyond the political domain.

    At its core lies Girard’s well-known insight: human desire is not autonomous. We do not desire objects directly; we desire them through others. Desire is mediated, structured by imitation, and therefore inherently relational. Truong’s article shows how this anthropological insight can be detached from its original moral and religious context and repurposed as an instrument of power. If human beings imitate, then whoever shapes the environment of imitation acquires a subtle but decisive form of influence.

    The article does not develop this point systematically. But it suggests something more general: ideas about human behaviour, once translated into systems, cease to be descriptive and become operative. They begin to organise reality.

    From Mimetic Desire to Delegated Thinking

    Girard’s theory is not simply a claim about desire; it is a claim about structure. Desire is mediated through a third element — a model whose desire we imitate — forming a triangular relation between subject, model and object. What appears as personal preference is, in fact, already shaped before it becomes conscious.

    The algorithmic environment introduces a structurally analogous mediation, but at a different level. It does not primarily organise what we desire, but how we think. Increasingly, we do not approach a problem directly. We approach it through systems that suggest formulations, generate reasoning paths, and prestructure the space of possible answers.

    The parallel is not rhetorical. This analogy, however, must be handled with precision. Mimetic desire, in Girard’s account, is constitutive of human subjectivity, whereas delegated thinking is a historically contingent tendency enabled by specific technological environments. The parallel does not collapse the two; it clarifies how a structural feature of human behaviour can be intensified under new conditions.

    In both cases, orientation occurs before deliberation. In both cases, the mediation is largely invisible. And in both cases, the subject experiences the result as its own.

    This is what I would call delegated thinking: the progressive transfer of cognitive initiative — the act of formulating, structuring and orienting inquiry — from the individual to the system, not through explicit decision, but through habituation.

    We do not stop thinking. We stop initiating thought.

    Arendt’s Warning Revisited

    This shift is not merely cognitive. It has a moral dimension.

    In her analysis of totalitarianism, and most explicitly in her account of the trial of Adolf Eichmann, Hannah Arendt described a form of failure that was neither intellectual incapacity nor ideological fanaticism. It was a failure of judgement. Eichmann did not lack intelligence. He lacked the ability to step back from the framework within which he operated and to evaluate it.

    He functioned within a system that had already structured the space of possible decisions. His role was to execute, not to question. The language he used, the categories he employed, the problems he addressed — all were given.

    The parallel with our present situation must be handled with care. But structurally, the risk is recognisable. When systems begin to pre-formulate questions, suggest directions, and define relevance, the individual may retain activity while gradually losing authorship. One continues to operate, but within a space that one has not defined.

    Delegated thinking is not imposed. It is accepted because it is efficient.

    The Educational Risk

    The implications for education are therefore not superficial. They do not concern themselves only with the use of tools, but also with the formation of the mind.

    If cognitive initiative is progressively externalised, several shifts follow. Intellectual effort is reduced not because students are unwilling, but because the system removes the need to struggle. Correctness becomes detached from understanding, as valid outputs can be produced without reconstructing the reasoning that justifies them. Authorship weakens because the origin of ideas becomes increasingly opaque.

    But the most significant shift is more subtle. Criteria themselves — what counts as a good question, a relevant argument, a meaningful problem — begin to migrate from the individual to the system. Traditionally, these criteria are not given in advance. They are formed through the very process of inquiry: by struggling with ambiguity, by testing formulations, by learning to distinguish what is trivial from what is meaningful. When systems pre-select directions, highlight what appears relevant, and suggest lines of reasoning, this formative process is short-circuited. The student no longer develops an internal sense of orientation. Instead, orientation is received.

    What is lost is not information. It is the capacity to recognise significance.

    A Necessary Clarification

    The value of the diagnosis suggested in Truong’s article lies in reminding us of something essential: human beings are not isolated rational agents. We are shaped by structures we do not fully perceive. Artificial intelligence does not introduce this condition. It amplifies it.

    But amplification changes scale, and scale changes nature. When mediation becomes constant, immediate and invisible, its effects are no longer episodic. They become the environment itself.

    A Final Distinction

    The question is no longer whether students can think.

    It is whether they remain capable of initiating thought, or whether they limit themselves to operating within structures that have already defined what is worth asking, what counts as relevant, and which paths are available.

    And this precedes any discussion about correctness, performance, or even understanding.

    Because before any answer can be judged, something more fundamental must occur: the question must be posed.

    And the capacity to pose it — to define the problem rather than merely respond within it — is where judgement begins.

  • What Habermas Can Still Teach Us in the Age of Artificial Intelligence

    In memory of a philosopher who taught us how to think

    The death of the German philosopher Jürgen Habermas this week invites more than a moment of remembrance. It invites a question: What can his thought still teach us about the technological moment we are living through?

    Habermas did not write about generative artificial intelligence. Yet one of the central distinctions in his philosophy helps illuminate an essential aspect of the age of intelligent machines. Not all forms of rationality are the same. Some forms of reasoning are oriented toward calculation. Others are oriented toward understanding. Artificial intelligence forces us to look again at that difference.

    The Expansion of Calculation

    Habermas described one form of reasoning as instrumental rationality. It is the kind of reasoning we use when we try to achieve a goal as efficiently as possible. It asks practical questions: what works? What produces the best outcome? How can a process be optimised? This logic has shaped much of modern society. It lies behind engineering, technological systems and many forms of scientific modelling.

    Artificial intelligence represents an extraordinary expansion of this capacity. Machines can process immense quantities of information, detect patterns and generate plausible responses in seconds. In that sense, artificial intelligence dramatically expands our ability to produce answers. But producing answers is not the same as understanding them.

    When Answers Come Before Understanding

    A system can now generate convincing explanations about history, economics or philosophy. It can construct arguments that appear coherent and even persuasive. Yet something important is missing.

    The machine does not understand the meaning of the words it produces. It does not grasp the implications of the argument it constructs. And it does not stand behind the claims it generates. In a strict sense, it produces answers without participating in understanding.

    Habermas described another form of rationality that belongs to human beings. He called it communicative rationality. This form of reasoning appears when people try to justify their claims to one another, explain their reasons and interpret meaning together. It is not primarily about efficiency. It is about making sense of things.

    An Older Philosophical Insight

    In this respect, Habermas stands within a much older philosophical tradition. Already in antiquity, Aristotle distinguished between different forms of knowledge. Some forms of knowledge are technical. They concern procedures, methods and the production of results. But Aristotle also described another form of intelligence. He called it phronesis, practical wisdom — the ability to judge well in situations where rules alone are not enough.

    Around the same time, Plato described education in a similar spirit. In the famous allegory of the cave, learning is not the simple accumulation of information. It is a turning of the mind toward reality. The Greeks had a word for this intellectual shift: periagoge — a turning of the soul.

    Education, in this sense, is not simply about receiving answers. It is about learning to see differently.

    The Space of Judgment

    This is where judgment appears. Judgment requires interpretation. It involves weighing arguments, recognising uncertainty and deciding what makes sense in a concrete situation.

    Artificial intelligence can help us explore possibilities. It can suggest explanations, detect patterns and generate answers we might not have considered. But it cannot decide what those answers ultimately mean. It cannot justify them in a public conversation. And it cannot assume responsibility for their consequences.

    Those tasks remain human.

    Artificial intelligence expands our capacity to generate answers.

    Education must therefore help students undergo a kind of intellectual periagoge — a turning of the mind from searching for answers to exercising judgment about them. Artificial intelligence may generate answers, but education exists to form the minds capable of understanding and judging them.

  • AI and the Disappearance of Intellectual Silence

    The Silence Before Thought

    Much intellectual work begins in silence.

    Before an argument is written, before a concept is articulated, there is usually a period in which the thinker does not yet know exactly what they think. The page remains blank, the question remains unresolved, and the mind explores possibilities that have not yet taken linguistic form.

    This moment can feel uncomfortable. It is the experience of hesitation, uncertainty, or partial intuition. Yet it often plays a crucial role in the formation of thought. During that interval, ideas reorganise themselves silently: connections appear, assumptions are questioned, and fragments of understanding begin to cohere into a structure.

    For centuries, intellectual practice has included this phase. Philosophical reflection, scientific discovery and creative writing all depend on moments in which answers are not immediately available.

    Thought begins in that space.

    The End of the Blank Page

    Artificial intelligence changes this situation in subtle yet profound ways.

    For the first time, when uncertainty appears, a machine can immediately generate a response. The page no longer needs to remain blank. A plausible explanation, an argument, a draft or a solution can appear almost instantly.

    The disappearance of the blank page may appear to be a purely practical advantage.. Writers overcome initial hesitation. Students obtain examples of how an argument might be structured. Researchers can quickly explore possible lines of reasoning.

    Yet something deeper may be changing as well. The silence that traditionally preceded thought can now be bypassed.

    When Answers Arrive Too Quickly

    The problem is not that artificial intelligence produces answers. Many of those answers are useful starting points for reflection. The difficulty arises when the appearance of an answer replaces the experience of thinking.

    If a student confronts a difficult question and immediately receives a well-formed response generated by an AI system, the intellectual process that would normally unfold in the absence of that answer may never take place. The learner moves directly from question to result, without inhabiting the intermediate space in which uncertainty forces reflection.

    The page is filled. But the mind may not yet have begun to work.

    Two Ways of Using Artificial Intelligence

    This does not mean that artificial intelligence necessarily eliminates intellectual silence. In practice, two different patterns of use are emerging.

    In the first, AI replaces the moment of reflection. The user accepts the generated answer as a substitute for thinking through the problem. The machine provides language, structure and reasoning that the user does not attempt to reconstruct.

    In the second, AI becomes part of a dialogue. The initial intuition comes from the user, and the system helps develop, test or articulate that intuition. The silence preceding the idea still exists, but the elaboration of the idea becomes interactive.

    The difference is subtle but decisive for intellectual formation. In the first case, thinking is delegated. In the second, thinking is extended.

    The Educational Risk

    From an educational perspective, the risk is not simply that students will use artificial intelligence to complete assignments. The bigger risk is that they may never experience the silence that precedes genuine thinking.

    That silence often contains the most formative moment in learning: the moment in which the student does not yet know how to proceed and must begin to organise their own reasoning. It is in that space that questions emerge, hypotheses are tested, and conceptual structures begin to appear.

    If answers always arrive before that moment unfolds, something essential to intellectual formation may be lost. Learning may become faster. But thinking may become thinner.

    Silence and Questions

    Artificial intelligence may therefore transform not only how answers are produced, but where silence appears in intellectual life.

    If machines increasingly generate responses, the crucial silence may no longer occur before the answer, but before the question. In a world where answers are abundant, the scarce intellectual capacity may be the ability to formulate questions that genuinely matter.

    Thinking has never consisted only in producing answers. It also consists of inhabiting the silence from which meaningful questions emerge.

    Further reflections

    This essay connects with:

    • The Renaissance Workshop and the Limits of Artificial Intelligence

    • Artificial Intelligence as a Stress Test for Critical Thinking

    • When the Answer Is Correct but the Mind Is Not Formed

  • Thinking with Others. On Intellectual Debts in an Age of Intelligent Machines

    Thinking in Conversation

    The reflections developed in these essays did not emerge in isolation. They grew out of teaching, reading, conversation and reflection on how intelligent technologies are reshaping the conditions under which learning and judgment take place.

    Writing these essays has made something increasingly clear to me. The questions raised by artificial intelligence in education do not arise in an intellectual vacuum. They often echo problems that earlier thinkers had already explored in different contexts.

    What I am attempting here is therefore not academic philosophy. It is an effort to think about education at a moment of technological transformation while remaining attentive to thinkers who have examined the nature of knowledge, judgment, understanding, and human action.

    In that sense, these reflections are not written alone. They take place within a much longer conversation.

    Execution and Judgment

    One of the central distinctions in my work concerns the difference between execution and judgment.

    Artificial intelligence is increasingly capable of executing tasks that once required significant human expertise. It can produce answers, generate explanations, analyse data and structure arguments with remarkable speed.

    Yet producing an answer is not the same as deciding whether that answer should be trusted, interpreted or applied.

    That second activity requires judgment.

    The idea that judgment represents a distinct intellectual capacity has deep philosophical roots. Aristotle described phronesis, often translated as practical wisdom, as the ability to deliberate well about what should be done in concrete situations. It is not merely technical knowledge but a form of reasoning that weighs circumstances, consequences and responsibilities.

    Centuries later, Immanuel Kant described judgment as the faculty that allows us to apply general principles to particular situations when no rule can determine the answer in advance. Judgment becomes necessary precisely when the mechanical application of rules is insufficient.

    These traditions matter because judgment is not only about solving intellectual problems. It is also about orienting action. In that sense, it belongs to the conditions of freedom. A person who cannot judge may still execute tasks, comply with instructions or perform competently. But they cannot fully decide how to act, what to value or what responsibility to assume.

    These insights acquire renewed relevance in the age of intelligent machines. Artificial intelligence can increasingly perform tasks of execution. The human challenge lies in cultivating the capacity to judge when and how those outputs should be used—and, more deeply, how one should live and act in a world increasingly shaped by them.

    Formation and Results

    A second distinction that runs through these essays concerns the difference between obtaining correct results and undergoing intellectual formation.

    Modern educational systems often rely on tasks whose success can be measured through correct outputs. Students solve equations, produce essays or summarise texts. These results are frequently treated as evidence of understanding.

    Yet the process through which understanding develops is not identical to the result itself.

    This insight also has deep philosophical roots. In the Phenomenology of Spirit, Hegel emphasised that truth emerges through a process in which consciousness confronts contradictions and gradually transforms its understanding. Knowledge cannot be separated from the path through which it is achieved.

    In educational thought, John Dewey similarly insisted that learning is not the passive acquisition of information but the result of experiences that reshape how individuals perceive and interpret the world.

    Artificial intelligence introduces a new tension here. When correct answers become instantly available, it becomes possible to obtain results without undergoing the intellectual processes that form the thinker.

    The answer may exist. But the mind may not have been formed.

    This distinction matters because education is not only about enabling performance. It is about shaping the kind of person who will later judge, decide and assume responsibility. Formation is what makes freedom intellectually possible.

    Technologies That Reshape Thought

    A third line of reflection concerns the role of technologies in shaping thought habits.

    The idea that intellectual tools influence how human beings think is not new. Plato already raised this concern when reflecting on writing in the dialogue Phaedrus. Writing, he suggested, might expand the transmission of knowledge while simultaneously weakening memory and the internal processes of reflection.

    Much later, scholars such as Walter Ong and Neil Postman examined how different communication technologies reshape the structures of thought and cultural life. More recently, Nicholas Carr has argued that digital environments may alter attention and patterns of cognitive engagement.

    Artificial intelligence represents a new stage in this long history of cognitive technologies. Unlike previous tools, generative systems do not merely store or transmit knowledge. They actively produce language, explanations and interpretations.

    As a result, they increasingly participate in the very processes through which ideas are articulated and understood.

    Understanding how these technologies reshape intellectual habits, therefore, becomes a central educational question. If technologies help configure the ways in which we attend, interpret and respond, they also influence the conditions under which judgment and freedom are exercised.

    The Conditions of Thinking

    A further theme concerns the conditions under which thinking itself becomes possible.

    Philosophers and cultural critics have often emphasised the importance of moments of uncertainty, difficulty or silence in the development of thought. Reflection frequently begins not with immediate answers but with questions that remain unresolved for a time.

    Generative artificial intelligence alters this condition. The blank page, the moment of hesitation and the temporary absence of answers may increasingly be replaced by instant responses.

    Yet the space in which thought emerges may depend precisely on those moments in which answers are not immediately available.

    Recognising this tension does not require rejecting new technologies. It requires understanding how they reshape the intellectual environments in which learning takes place.

    A person who never inhabits uncertainty may become more efficient, but not necessarily more capable of judgment. And without judgment, freedom itself becomes thinner: one responds more quickly, but chooses less deeply.

    Thinking with Others

    Looking back at the essays collected on this site, I have gradually realised that many of the ideas they contain belong to a much older conversation.

    Aristotle reflected on practical judgment. Kant examined the faculty that allows us to apply principles to particular cases. Hegel explored the formative process through which understanding develops. Dewey examined learning as experience. Plato, Ong, Postman and others considered how technologies transform the conditions of thought.

    To this conversation, I would also add those thinkers who help us see that judgment matters not only because it improves decisions, but because it belongs to a fully human life. Sartre, for example, reminds us that human beings cannot escape responsibility for their choices. Even when systems produce recommendations, someone must still decide whether to adopt them and answer for the consequences.

    None of these thinkers addressed artificial intelligence in education as we encounter it today.

    Yet their work illuminates the questions we now face.

    My intention in writing these essays is not to produce academic philosophy, but to think carefully about education at a moment when intelligent technologies are rapidly transforming how knowledge is produced, communicated and applied.

    If these reflections contribute something of value, it is because they take place within this wider tradition of thinking about how human beings learn, judge, act and live freely.

    In that sense, the work presented here is not written alone.

    It is written in conversation with those who have long reflected on how human beings learn, judge and act in the world.

  • Thinking When Machines Think

    Education and Human Judgment in the Age of Artificial Intelligence

    Artificial intelligence changes not only what we can do, but also how we learn to think. The central task of education is therefore the formation of judgement.

    For centuries, education prepared individuals to execute intellectual tasks: solving equations, analysing evidence, drafting arguments, diagnosing problems and producing solutions. Mastery meant learning how to apply knowledge correctly and efficiently, and professional expertise was largely defined by the capacity to perform complex cognitive work reliably.

    Today, that assumption is beginning to change. Systems capable of generating explanations, analysing data, producing text and assisting decision-making are rapidly becoming part of our intellectual environment. Tasks that once required years of training can now increasingly be initiated — and sometimes completed — by machines.

    Much of the discussion surrounding this transformation focuses on productivity, efficiency or technological capability. Yet the deeper question raised by artificial intelligence is not primarily technological. It is educational.

    If intelligent systems begin to perform many of the cognitive tasks once associated with expertise, what then becomes the purpose of education?

    For much of the modern era, education has trained individuals to execute. Students learned procedures, mastered analytical techniques and acquired the intellectual tools necessary to produce correct results. Execution became the foundation of professional competence and the measure of intellectual achievement.

    Artificial intelligence is now rapidly expanding the automation of execution. Machines can assist with tasks that require defined expertise: analysing evidence, writing reports, generating explanations and proposing solutions. This development does not make human knowledge irrelevant. On the contrary, it clarifies the importance of a different human capacity.

    Judgement is not simply the production of an answer. It is the capacity to determine which answers matter, why they matter and whether they should be trusted. It involves recognising context, weighing competing considerations, anticipating consequences and assuming responsibility for decisions. In a world where machines increasingly produce outputs, the value of judgement becomes clearer.

    Results and Formation

    Artificial intelligence introduces another educational challenge: the distinction between obtaining results and being intellectually formed through the process of learning.

    Correct answers have never been the ultimate goal of education. Learning involves intellectual struggle. Students encounter difficulty, confront contradictions, revise arguments and gradually reconstruct their understanding. Through this process, knowledge becomes internal rather than external. The learner is transformed by the act of thinking.

    Artificial intelligence makes it possible to obtain results without passing through that formative process. A student can request a complete mathematical solution, a structured essay or a plausible explanation within seconds.

    The output may be correct. Yet if the reasoning behind it has not been reconstructed by the learner, something essential may be missing. The student possesses the answer but has not undergone the thinking-through that makes understanding genuine.

    This distinction between result and formation becomes one of the central educational questions of the age of intelligent machines.

    Language and Thought

    Another dimension of this transformation concerns language itself.

    Language is not merely a tool for expressing thought; it shapes the categories through which we understand the world. Philosophers such as Ludwig Wittgenstein emphasised that the limits of our language shape the limits of our thinking.

    Artificial intelligence increasingly generates the language that circulates through our intellectual environments: summaries, explanations, reports, analyses and narratives. As algorithms produce a growing proportion of the texts we read and write, they inevitably begin to influence how ideas are framed and understood.

    Education must therefore cultivate not only knowledge but awareness of the linguistic frameworks through which knowledge appears.

    Seeing Structure

    Human understanding does not arise from the accumulation of isolated fragments of information. Gestalt psychology showed that the mind perceives meaningful structures rather than disconnected elements. Understanding occurs when relationships between ideas become visible, and concepts begin to form coherent patterns.

    Artificial intelligence excels at producing informational fragments: explanations, examples, summaries and answers. But education must ensure that learners still develop the ability to perceive the conceptual structures that connect those fragments.

    Without that capacity, knowledge risks becoming a collection of correct statements without a coherent understanding.

    An Ancient Question

    Concerns about intellectual technologies are not new.

    In the dialogue Phaedrus, Plato recounts a myth in which the Egyptian god Thoth presents writing as a gift that will improve wisdom and memory. The response is sceptical. Writing, the king warns, may create the appearance of wisdom without genuine understanding:

    “You provide your students with the appearance of wisdom, not with its reality.”

    Plato feared that writing might allow people to possess knowledge without truly understanding it. Artificial intelligence raises a similar question today. When answers become instantly available, education must ensure that the intellectual processes through which understanding develops are not lost.

    The Educational Task Ahead

    Artificial intelligence will undoubtedly continue to transform how knowledge is produced, distributed and applied. The question is not whether education should use these tools. It must.

    The question is: what kind of minds should education form in a world where machines can increasingly execute cognitive tasks?

    The answer lies in cultivating intellectual capacities that cannot be automated: discerning what matters, recognising structure and context, weighing competing considerations and assuming responsibility for decisions. These capacities define judgement.

    In the age of intelligent machines, the central task of education is not to compete with machines in execution, but to form individuals capable of governing the systems they use. Technology expands what we can calculate and create, but it also transforms the intellectual environment in which thinking develops.

    Execution can be delegated.

    Judgement cannot.

    And preserving our capacity to judge — to weigh reasons carefully, determine what deserves priority and assume responsibility for our conclusions — may therefore become one of the defining responsibilities of education in the decades ahead.

  • When the Answer Is Correct but the Mind Is Not Formed

    Result, Formation, and the Risk of Intellectual Alienation in the Age of AI

    Artificial intelligence can generate correct answers with remarkable speed. But education has never been about correct outputs alone. It concerns formation. If students bypass error, contradiction and reconstruction, they may obtain results without being transformed by the process of thinking. In the age of intelligent machines, the risk is subtle: not technological domination, but intellectual alienation.

    I am not a Hegel specialist. What I retain from university lectures is less doctrinal precision than a strong intuition: knowledge, for Hegel, was not merely a correct conclusion. It was a movement. Consciousness develops by encountering its own limits, by confronting contradiction, by being forced to revise itself.

    Truth, in that sense, is not static. It is developmental. That intuition acquires renewed relevance today.

    The Risk of Confusing Result with Formation

    Artificial intelligence produces results — often highly competent ones.

    Examine these situations

    • A student can obtain a fully worked-out mathematical proof in seconds-
    • An essay can be structured with logical coherence and sophisticated vocabulary.
    • A philosophical argument can be drafted with plausible references and elegant phrasing.

    From the perspective of output, nothing appears missing. But something essential may be absent.

    If a student obtains a correct solution generated by AI without having passed through the process of uncertainty, error, contradiction and reconstruction, they have obtained a result — but not formation. The distinction is not rhetorical. It is structural.

    Consider a mathematics classroom. A student submits a flawless solution to a complex problem. When asked to explain why a particular step was necessary, they hesitate. They cannot retrace the reasoning without reopening the tool that generated it. The answer is correct. The reasoning is not owned.

    Or consider an essay discussion.

    The written argument is coherent and refined. Yet when pressed to defend a claim or respond to an objection, the student struggles. The language was produced. The intellectual struggle was not. Again: result without formation.

    Education has always aimed at something deeper than correct performance. It seeks the capacity to think through difficulty, to recognise contradictions, to reformulate positions and to assume responsibility for conclusions. Artificial intelligence makes it easier than ever to bypass that formative arc.

    Why Struggle Matters

    In Hegelian terms — as I understand them — development occurs through tension. Each stage of understanding reveals its limits. Failure is not accidental; it is formative. It compels reconfiguration.

    Experience is not additive. It is transformative. Educationally, this means:

    • A student who wrestles with a problem and revises their understanding is changed by that struggle.
    • A student who reconstructs an argument after critique expands their horizon.
    • A learner who confronts the limits of their reasoning develops judgement.

    Remove the friction, and you remove the transformation. If AI eliminates confusion, hesitation, and the discomfort of not knowing where to begin, something essential is lost.

    Not efficiency. Formation.

    The Subtle Emergence of Intellectual Alienation

    There is a bigger risk: intellectual alienation.

    Alienation occurs when what we produce no longer feels like something we have genuinely thought.

    • A student submits polished work they cannot defend.
    • A professional circulates an AI-generated report they cannot fully explain.
    • An institution adopts algorithmic metrics whose logic no one has critically examined.

    In each case, the subject becomes separated from their own output. The work is impressive. Ownership is thin.

    This is not technological domination. It is voluntary cognitive delegation — the quiet outsourcing of the very processes that form judgement.

    And it happens gradually:

    • Accepting a generated explanation without retracing the reasoning.
    • Editing AI output instead of drafting from conceptual clarity.
    • Prioritising speed over intellectual integration.

    Over time, habits shift. The student becomes a curator of answers rather than a constructor of thought. The professional becomes a supervisor of outputs rather than a judge.

    Practical Implications for Education

    The response is not a prohibition. It is design.

    Education must structure learning so that formation cannot be outsourced.

    This may involve:

    • Requiring oral defence of written work.
    • Asking students to reconstruct reasoning without technological assistance.
    • Evaluating explanation as carefully as correctness.
    • Designing assessments that value revision history and reflective commentary.
    • Introducing deliberate moments of contradiction and challenge.

    In mathematics, this means asking not only for the derivation, but for an account of why each step is necessary.

    In literature, it means defending interpretative choices under questioning.

    In philosophy, it means engaging objections rather than merely presenting positions.

    In every discipline, it means preserving the movement from uncertainty to clarity.

    A Question Worth Asking

    Artificial intelligence intensifies an old philosophical question: Is knowledge something we possess as a product, or something we become through experience?

    If education delegates formative struggle to AI, it risks producing technically fluent but judgment-poor graduates. They may generate outputs efficiently.

    But they may not have been transformed by the thinking.

    In the age of intelligent machines, the decisive issue is not whether answers are correct. It is whether the person behind them has been formed.

    Execution can be delegated.

    Judgement cannot.

    If education forgets that distinction, the cost will not be technological. It will be anthropological.

    This essay connects with:

  • AI, Plagiarism, and Intellectual Authorship in the IB Diploma Programme

    Clarifying Authorship and Responsibility in an Age of Intelligent Tools

    Artificial intelligence has entered education not with the gradual discretion of earlier technologies but with striking speed. Nowhere is the resulting uncertainty more visible than in academically demanding contexts such as the International Baccalaureate Diploma Programme. Teachers, examiners, and students increasingly ask a deceptively simple question: when a student uses AI to prepare an Extended Essay, a Theory of Knowledge essay, or an Internal Assessment, is that plagiarism?

    The honest answer resists simplification. AI use is not automatically plagiarism, yet neither can it be considered neutral. What is required is not prohibition born of anxiety, but a clearer understanding of authorship, intellectual responsibility, and the purpose of academic work itself.

    Plagiarism, in its classical sense, involves presenting identifiable words or ideas belonging to another author without acknowledgement. Generative systems complicate this definition because they rarely reproduce specific texts verbatim. Instead, they produce new linguistic formulations shaped by patterns within patterns. The result may sound authoritative, but it is not normally traceable to a single identifiable source.

    If a student uses AI to clarify expression, test an emerging idea, organise an argument, or overcome linguistic insecurity, the situation resembles forms of academic mediation long accepted in education: editorial guidance, discussion with teachers, even sophisticated grammar tools. None of these, in themselves, negate authorship. Intellectual ownership rests less in phrasing than in the origin and defence of the thinking.

    In the context of the IB Diploma Programme — where independent inquiry and reflective judgement are central — this distinction acquires particular significance. If the research question, interpretative stance, and argumentative decisions are genuinely the student’s, AI functions as an instrument rather than an author. The decisive issue is not who produced the sentences, but who assumes responsibility for the ideas.

    Yet the boundary cannot be ignored. When AI generates analysis or interpretation that the student neither understands nor critically assesses, the educational process is hollowed out. Even if the wording is technically original, intellectual labour has effectively been outsourced. That situation may not always fit a narrow legal definition of plagiarism, but academically, it risks compromising integrity.

    Transparency, therefore, becomes essential. Concealing substantial AI assistance undermines the trust upon which academic evaluation depends. Equally significant is reliability. Generative systems can produce inaccuracies, invented references, or oversimplified interpretations. Incorporating such material uncritically does not necessarily constitute plagiarism, but it falls short of scholarly responsibility. Verification, discernment, and intellectual caution remain indispensable academic virtues.

    Much of the anxiety surrounding AI reflects legitimate concerns, though it sometimes overshoots its target. The assumption that AI-assisted writing cannot represent authentic student work presupposes a rather romantic image of solitary authorship that has rarely existed in academic practice. Writing has always involved dialogue — with teachers, peers, editors, and prior texts. Tools evolve; the social nature of knowledge does not.

    The deeper issue, therefore, is not technological but educational. If academic tasks reward procedural execution alone, delegation becomes tempting. If they require judgment — the capacity to weigh evidence, defend interpretation, and assume responsibility for conclusions — authorship remains irreducibly human.

    Clear expectations do not constrain students; they protect their intellectual development. Conceptual clarity, however, must translate into operational judgement.

    Questions That May Help Clarify Responsible AI Use

    The following questions are not intended as regulation, but as prompts for judgment. They may help students, teachers and schools reflect on whether AI is supporting learning — or quietly replacing it.

    • Does the research question and central line of argument genuinely originate from the student?
    • Can the student clearly explain and defend every major claim without relying on the AI tool?
    • Has AI been used primarily to clarify language, refine structure or test emerging ideas, rather than to generate substantive analysis?
    • Did the student critically evaluate, revise or reject suggestions produced by the tool?
    • Have all references, data and quotations been independently verified using reliable sources?
    • Has engagement with required primary and secondary sources been maintained?
    • Has any significant AI assistance been transparently acknowledged?
    • If asked to discuss the work orally, without access to AI, would the student demonstrate clear understanding and independent reasoning?
    • Would the quality of thinking — even if not the polish of expression — remain recognisably their own?

    A final reflection may be decisive.

    If artificial intelligence is present in academic work, the essential question is not whether a tool has been used, but what kind of thinking has taken place.

    That question cannot be resolved by detection software or automatic rules. It requires judgement — exercised by students, teachers and institutions alike.

    A broader philosophical exploration of judgment and authorship in the age of intelligent tools is developed in The Renaissance Workshop and Human Judgment in the Age of Artificial Intelligence.