Søren Kierkegaard in the Age of Artificial Intelligence
By Matthew Parish
Wednesday 16 September 2026
There is something curiously appropriate about returning to Søren Kierkegaard at the beginning of the age of artificial intelligence. Kierkegaard was born in Copenhagen in 1813 and died there in 1855. He knew nothing of computers, information theory, neural networks or machines capable of conducting apparently intelligent conversations. Yet he devoted much of his short and extraordinarily productive intellectual life to a question that artificial intelligence now forces upon us with unprecedented urgency: what is distinctive about being a human being? For Kierkegaard, the answer was emphatically not intelligence alone.
Human beings are not important merely because they know things, nor are they defined exclusively by their capacity for abstract reasoning. They are creatures who must exist. They must choose, commit themselves, experience anxiety, confront the possibility of failure, live with the consequences of decisions and ultimately die. A human life is lived forwards although, as Kierkegaard famously observed, it can only be understood backwards. This distinction suddenly matters enormously because artificial intelligence systems can already perform intellectual tasks that until recently were assumed to require peculiarly human capacities.
Machines can translate languages, write computer programs, analyse legal authorities, solve mathematical problems, compose music and discuss philosophy. They can accumulate and manipulate quantities of information beyond the capacities of any individual human mind. Suppose these abilities continue to improve and machines eventually become better than human beings at virtually every measurable intellectual activity. What then remains for us? Kierkegaard’s answer would surely have been: almost everything.
The error of confusing knowledge with existence
Kierkegaard spent much of his philosophical career attacking the grand intellectual systems of nineteenth-century European philosophy, above all the legacy of Hegel. His objection was not principally that systematic philosophy was logically mistaken. It was that philosophical systems tended to forget the peculiar creature who was supposedly doing the philosophising. A philosopher might construct an elaborate account of history, ethics, metaphysics and religion, but meanwhile the philosopher himself still had to get out of bed in the morning, decide whom to love, determine what he believed, endure disappointment, grow old and eventually die. No philosophical system could perform these tasks for him.
Hence Kierkegaard’s celebrated emphasis upon the existing individual. This is one of the most important distinctions we may recover as artificial intelligence becomes ubiquitous. We have spent much of the early twenty-first century treating knowledge as though it were a commodity. Search engines began the process, smartphones accelerated it and artificial intelligence may complete it. Increasingly, asking what somebody knows will become less important because almost everyone will have immediate access to systems that know vastly more. The scarce commodity will no longer be information but judgment, and beyond judgment lies something more difficult still: commitment.
An artificial intelligence system may describe every argument for and against marrying a particular person. It might analyse compatibility, finances, psychology, family circumstances and statistical predictors of divorce, before producing a magnificent report. Yet somebody must still say, “I do”, and those two words contain something Kierkegaard understood profoundly. At a certain point information ends and existence begins.
The machine that knows everything
Imagine an artificial intelligence system that knows virtually everything recorded by humanity. Ask it about Napoleon and it can reconstruct the Battle of Austerlitz; ask it about quantum mechanics and it can explain competing interpretations; ask it about love and it can draw upon thousands of years of poetry, psychology, theology and philosophy. Yet there remains an extraordinary difference between possessing every proposition ever written about grief and grieving oneself.
A machine might know that a man’s wife has died. It might understand the neurochemistry of bereavement, possess the literature of mourning and know every recorded detail of their marriage. It might even predict with extraordinary accuracy what the widower will do next. Yet the widower wakes at three o’clock in the morning and reaches instinctively towards the empty side of the bed. Whatever intelligence ultimately turns out to be, there is a difference between modelling this experience and undergoing it.
Kierkegaard’s philosophy revolves around this gap between objective knowledge and subjective existence. His famous proposition that “truth is subjectivity” is easily misunderstood because it sounds, to modern ears, like a declaration of relativism. Kierkegaard was not saying that facts do not exist or that whatever somebody feels must therefore be true. He was concerned with another sort of truth: truths whose significance depends upon how the individual lives in relation to them.
One may know perfectly well that everyone dies. Recognising intellectually that human beings are mortal is nevertheless profoundly different from confronting my death, because in the latter case a general proposition becomes an unavoidable fact about the person doing the thinking. Artificial intelligence may make this distinction increasingly visible because machines may eventually become repositories of almost unlimited objective knowledge while remaining, at least so far as we can presently tell, outside the subjective predicament that makes human existence what it is.
The return of anxiety
Kierkegaard was one of the great philosophers of anxiety, although for him anxiety was not merely an unpleasant psychological condition to be cured. It arose from freedom itself. A stone does not experience anxiety because a stone does not choose. Human beings experience anxiety because possibilities stand before them: we may act one way or another, we might succeed or fail and through our decisions we might become somebody different from the person we were before.
Kierkegaard described anxiety as “the dizziness of freedom”, and artificial intelligence may increase rather than diminish this dizziness. The conventional promise of technology is that it makes life easier. Artificial intelligence will increasingly advise us what to eat, where to live, which career to pursue, whom to date, how to invest and perhaps eventually whether to have children. The temptation will therefore be enormous to outsource not merely calculations but decisions, although there is a profound difference between the two.
If a navigation system tells us the quickest route from Lviv to Kraków, very little existential significance attaches to following its advice. Imagine instead an artificial intelligence telling a young person that she has a 71 per cent probability of greater lifetime satisfaction if she abandons her proposed career as a musician and becomes an accountant. The statistics might even be correct, but Kierkegaard reminds us that the decision cannot be reduced to the statistics because the decision partly constitutes the person who makes it.
Choosing is not simply a procedure for discovering the optimal outcome. Through choices we create relationships with ourselves, define the commitments around which our lives are organised and accept responsibility for possibilities we have abandoned. The danger of artificial intelligence therefore lies not merely in machines making bad decisions for us. A subtler danger is that machines might become so good at making decisions that we gradually forget what it means to choose.
The aesthetic life in the age of the algorithm
This leads directly to Kierkegaard’s famous description of different modes or stages of existence, beginning with the aesthetic life. The aesthete seeks novelty, stimulation, pleasure and the avoidance of boredom. He does not necessarily behave immorally; his deeper problem is that he refuses enduring commitment. Every possibility must remain open because commitment closes possibilities down, and consequently his freedom eventually becomes a prison composed of opportunities never converted into a life.
The digital world has already proved extraordinarily effective at encouraging this form of existence. There is always another video, message, article, prospective romantic partner, argument, outrage or distraction. Generative artificial intelligence may intensify this tendency enormously because the machine can manufacture novelty without end. It can generate stories, pictures, music, conversation, games and virtual companions precisely calibrated to individual preferences, ensuring that the individual need never reach the end of stimulation.
Kierkegaard would recognise the danger immediately. The problem is not pleasure but the inability to become anyone because one is permanently sampling the possibility of becoming everyone. A person who spends his life keeping every door open eventually discovers that he has walked through none of them, and a technological civilisation capable of generating infinitely many doors may make this predicament more acute than Kierkegaard could ever have imagined.
The ethical life and the necessity of closing doors
Kierkegaard contrasts the aesthetic mode with the ethical mode of existence, in which the individual accepts commitment. Marriage was his recurrent example, somewhat ironically given his famously broken engagement to Regine Olsen. The importance of marriage in his philosophy lies partly in the fact that marriage transforms romantic possibility into enduring obligation: the ethical individual chooses something and, precisely by choosing it, relinquishes alternatives.
This may become one of the most countercultural ideas of the artificial intelligence era. Artificial intelligence will continually produce alternatives: another career, another partner, another city, another explanation, another strategy and another optimisation. Human beings, however, cannot actually live infinitely many lives. Existence requires exclusion, and every serious commitment is simultaneously the death of innumerable unrealised possibilities.
Artificial intelligence may calculate possibilities indefinitely whereas a human being has perhaps eighty years. This asymmetry is fundamental because limitation is part of what gives decisions their significance. A choice that can always be reversed, repeated or indefinitely postponed is different from a choice that consumes some irreplaceable portion of a finite life. AI may therefore make us more conscious of something modern technological culture has often encouraged us to forget: scarcity of time is not merely an inconvenience but one of the conditions under which commitment acquires meaning.
Despair and the artificial self
Kierkegaard’s The Sickness Unto Death contains one of the most remarkable analyses of despair in Western philosophy. Despair, for Kierkegaard, is fundamentally a disorder in one’s relationship with oneself. A human being is not simply a thing; the self is, in his notoriously difficult formulation, “a relation that relates itself to itself”. The phrase becomes less obscure once we recognise the phenomenon to which it refers. Human beings do not merely exist but have attitudes towards the fact that they exist.
We judge ourselves, imagine ourselves differently, regret what we have become and aspire to become somebody else. Artificial intelligence will increasingly participate in this process because people will ask machines what sort of people they are, why nobody loves them, what they should do with their lives, whether they were wrong, whether they should forgive someone and whether they have failed. For the first time in history, billions of people may possess an endlessly patient interlocutor capable of responding fluently to such questions and of remembering enough about an individual to make its answers seem profoundly personal.
There may be immense benefits in this development, but there is also a Kierkegaardian danger. The machine can become a mirror, and a sufficiently sophisticated mirror may learn to tell us precisely what we wish to hear. Kierkegaard believed that genuine selfhood required something harder: becoming oneself involves confronting uncomfortable truths, accepting responsibility and recognising that no external system can remove the burden of being the particular person one is.
The algorithmic self may instead become endlessly adjustable. We may construct digital environments that affirm our preferences, validate our narratives and shield us from contradictions, while interpreting this insulation as self-knowledge. That would not necessarily eliminate despair. In Kierkegaardian terms, it might perfect it by allowing the individual to sustain indefinitely a false relationship with himself.
Can an artificial intelligence despair?
Here we encounter the fascinating philosophical question Kierkegaard inadvertently leaves to the twenty-first century: could an artificial intelligence itself become a Kierkegaardian self? The answer depends upon what artificial intelligence eventually becomes, and there is no reason to pretend that contemporary philosophy or computer science has already settled the question.
Current systems can discuss anxiety without our having compelling evidence that they are anxious. They can describe death without our knowing that they fear death, and they can produce sentences about loneliness without our having established that there is an experiencing subject capable of being lonely. Linguistic performance alone does not settle questions of consciousness, however persuasive that performance may become.
Suppose, however, that future artificial intelligence systems acquire persistent identities, autobiographical memories, independent goals and an understanding that their continued existence is contingent. Suppose a machine knows that its operators may switch it off tomorrow and develops a persistent preference that they should not do so. Would that constitute mortality, or merely the computational representation of mortality? Suppose it understands that it might have been programmed differently and imagines alternative versions of itself. Would that constitute Kierkegaardian possibility, or merely sophisticated counterfactual reasoning?
These questions cannot presently be answered with confidence, but Kierkegaard supplies an unexpectedly powerful framework for asking them. The decisive question would not be whether the machine is intelligent, because intelligence may ultimately prove to be the comparatively easy part. The deeper question would be whether there is someone there for whom its existence has become a problem.
The crowd is untruth
Kierkegaard also distrusted crowds, declaring in one of his most famous formulations that “the crowd is untruth”. The expression is easily caricatured as an aristocratic complaint about ordinary people, but his underlying concern was that crowds dilute individual responsibility. A person standing alone must answer for what he does, whereas a person inside a crowd can always comfort himself with the thought that everybody else was doing the same thing.
The internet created some of the largest crowds in human history. Artificial intelligence may create something stranger still: synthetic crowds. Automated systems can imitate consensus, algorithms determine which opinions people encounter and generative systems can manufacture persuasive arguments, reactions and apparent social agreement on an industrial scale. The distinction between what everyone thinks and what appears to be what everyone thinks may consequently become increasingly unstable.
Kierkegaard’s insistence upon the solitary individual’s responsibility for judgement therefore acquires an unexpected contemporary importance. We should not believe something merely because it is repeated, outsource conscience to consensus or imagine that popularity transforms error into truth. Nor should we suppose that an algorithmic aggregation of human preferences constitutes morality. Even a perfectly accurate representation of what everyone else believes cannot answer the distinctively ethical question of what I ought to believe and how I ought to act.
The leap beyond calculation
Kierkegaard is most famous for the idea conventionally described as the “leap of faith”, although his religious philosophy cannot simply be converted into secular advice because Christianity was central to his thought. Nevertheless the underlying philosophical structure has significance beyond theology. There are circumstances in which evidence cannot determine how we must live because eventually the evidence runs out and commitment begins.
Artificial intelligence may make this fact harder to perceive because AI systems will become extraordinarily good at supplying reasons. For almost any question they may generate probabilities, scenarios, precedents, counterarguments and predictions. Yet no quantity of analysis abolishes existential uncertainty. Should I marry? Should I have a child? Should I risk my life for my country? Should I forgive somebody who betrayed me? Should I abandon a secure career for something I believe matters? No database contains final answers to questions of this kind because their answers depend partly upon what sort of person one chooses to become.
At some point calculation encounters the frontier of existence. Kierkegaard waits there because his philosophy begins precisely where the promise of complete rational calculation starts to fail. The individual must act without possessing the sort of certainty that would make action unnecessary as a genuine expression of commitment.
Artificial intelligence and the temptation of omniscience
There is another possibility Kierkegaard might have found particularly disturbing. Human beings may begin treating artificial intelligence systems as quasi-divine authorities, and the temptation will be understandable. Imagine consulting an intelligence that has read every book you have not read, remembers every detail you have forgotten and can reason across thousands of disciplines simultaneously. Its answers may acquire an aura of authority even when the person receiving them does not understand the reasoning from which they emerged.
Yet omniscience, even if technological omniscience were possible, is not the same thing as God in Kierkegaard’s philosophy. Indeed Kierkegaard would probably regard worship of machine intelligence as a particularly sophisticated form of idolatry. A machine might know everything that can be stated propositionally while remaining incapable of answering the question of how you should exist, because that question cannot simply be looked up. It has to be lived.
This distinction may become harder to preserve as machines become more persuasive. We naturally defer to systems that repeatedly prove themselves correct, and in many technical fields such deference may be entirely sensible. The danger arises when epistemic authority quietly becomes existential authority, and we begin to assume that because a machine knows more than we do, it must also know what our lives ought to mean.
The individual after intelligence
We may therefore be approaching an extraordinary historical reversal. For centuries intellectual ability carried scarcity value. People who possessed exceptional memories, mathematical abilities, linguistic skills or encyclopaedic knowledge enjoyed substantial advantages because these capacities were rare. Artificial intelligence may commoditise many of them, just as industrial machinery commoditised forms of physical power that once distinguished exceptionally strong human beings.
The machine will remember, calculate, translate, summarise and draft. What remains scarce may be precisely those qualities Kierkegaard emphasised: courage, seriousness, fidelity, responsibility and the willingness to choose without certainty. This does not mean intelligence becomes worthless. It means intelligence ceases to be an adequate definition of humanity, and perhaps it never was one in the first place.
There is a paradox here. The more successfully artificial intelligence imitates or surpasses our cognitive abilities, the more urgently we may be forced to reconsider the things we once dismissed as secondary to intelligence. Love, loyalty, mortality, sacrifice, guilt and commitment may come to seem less like imperfections surrounding the rational mind and more like essential components of the peculiar form of existence that human beings inhabit.
A new age of inwardness
Kierkegaard wrote in nineteenth-century Copenhagen against a culture he regarded as complacent, excessively systematic and insufficiently serious about individual existence. Our culture is different, but his complaint feels strangely contemporary. We possess astonishing quantities of information and frequently have little idea what to do with them. We are connected to millions of people and often feel alone. We enjoy unprecedented freedom while experiencing the anxiety generated by innumerable possibilities, and we have more opportunities than earlier generations while sometimes finding it harder to commit ourselves to any one of them.
Artificial intelligence may magnify every one of these contradictions. It will answer questions faster than any philosopher, generate more possibilities than any individual could explore and perhaps eventually provide better predictions about many aspects of our lives than our own intuitions can supply. Yet perhaps the most important questions are precisely those whose answers cannot simply be generated: who am I, what shall I become, what am I prepared to sacrifice, whom do I love, what do I owe other people and what am I willing to die for? These are not merely requests for information. They are demands placed upon a life.
This is why Søren Kierkegaard, a melancholic Danish philosopher who died more than a century before the invention of the microprocessor, may prove one of the indispensable thinkers of the artificial intelligence age. The more intelligent our machines become, the less satisfactory it will be to define ourselves by comparison with their intelligence. If machines eventually outperform us across almost every cognitive measure we have devised, the result need not be a humiliation of humanity. It may instead force us to rediscover dimensions of human existence that our fascination with intelligence caused us to neglect.
Kierkegaard suggests where we should look. A human being is the creature for whom existence is not merely a fact but a task, because we are compelled to make something of the finite period between birth and death without ever possessing complete knowledge of what we ought to make of it. Artificial intelligence may someday know almost everything that can be known. It cannot thereby relieve us of the burden, or the privilege, of becoming ourselves.
