Existential Meaning and Values

In the spring of 2025, a philosophy professor named Isaac Levy received an email from a former student. She was a working writer in her thirties, good at her craft, and she had spent the past year watching large language models draft, in seconds, prose that took her days. "I'm having trouble understanding why what I do matters anymore," she wrote. "AI can write better than I can, faster, and without existential angst. If intelligence and creativity aren't uniquely human, what are we for?"

Levy had taught existentialism for three decades. He knew the reflexive answers — that machines only imitate, that meaning is something you make rather than something you have. He typed them out. Then he deleted them, because they sounded, even to him, like a man reassuring himself. The honest reply was that he did not know, and that not knowing had stopped being an abstract seminar puzzle and become a weight his students were carrying to work every morning.

That weight has a name. A small but growing group of philosophers, psychologists, and technology researchers now use the phrase AI existential crisis to describe something specific — and it is worth being precise, because the phrase invites confusion. It does not mean the fear that AI might one day kill everyone; that is existential risk, a separate literature about extinction. The AI existential crisis is about meaning, not survival. It is the disturbance that arrives when a technology begins to outperform us in the exact domains — thinking, creating, reasoning, understanding — through which our species has always explained why it matters. You can keep your job, your income, and your health and still be caught in it. That is what distinguishes it from ordinary career anxiety or economic disruption. A laid-off factory worker fears for the rent. A radiologist watching an algorithm read scans more accurately than she can fears something harder to insure against: that the thing she is for has been done better by something that does not need to be anything at all.

Four pillars, pushed at once

To see why the disturbance runs deep, it helps to return to the philosophy Levy taught. Twentieth-century existentialism — Jean-Paul Sartre, Albert Camus, Martin Heidegger, and their inheritors — grew out of a single, bracing premise: the universe hands us no purpose, so we must author our own. That project rests on four load-bearing ideas. Freedom: we are not fixed by biology or fate; we define ourselves through choice. Authenticity: we own those choices and forge our own values rather than borrow them. Self-determination: meaning accrues from the projects we pursue and the commitments we keep across a life. Uniqueness: your existence is irreplaceable — only you can live it.

AI does not chip at these one at a time. It presses on all four simultaneously, and each through a different mechanism.

Freedom is narrowed by prediction. Recommendation systems, ranking algorithms, and predictive models increasingly shape what we see, buy, believe, and choose next. The choices remain ours in name, but they arrive pre-curated by systems that have learned our patterns more precisely than we know them ourselves. Radical freedom is hard to feel when a machine finished your sentence before you started it.

Authenticity is complicated by generation. When a writer's essay, a designer's concept, or a student's argument is produced partly or wholly by a model and then presented as an expression of self, the boundary between self-expression and algorithmic output dissolves. To "be true to yourself" you need a self that is cleanly yours. That premise is exactly what AI-assisted cognition muddies.

Self-determination is constrained by relevance-setting. Increasingly, AI-driven markets and tools help determine which skills pay, which work is valued, and which goals are reachable — effectively drawing the boundaries inside which human projects must operate. When the parameters of a meaningful life are set by systems optimizing for something other than your flourishing, "determining yourself" becomes a smaller act than it sounds.

Uniqueness is undermined by replication. A model can absorb your writing style and produce more of it; it can mimic a painter's hand or a composer's ear. The irreplaceable begins to look reproducible. Existentialism assumed humans were the sole engines of meaning-making. When machines generate what looks like meaning — art, argument, consolation, insight — the tradition confronts a question it was never built to answer.

Meaning has more than one root, and AI reaches for each

Viktor Frankl, the psychiatrist who survived Auschwitz and wrote Man's Search for Meaning (1946), argued that human beings are driven by a "will to meaning," and that its frustration produces a distinctive kind of suffering he called the "existential vacuum." Meaning, in his account and in later research, is not monolithic. It grows from several roots. AI reaches for four of them at once.

Work is the most reliable. A job supplies not only money but identity, structure, social belonging, and the felt sense that one is good at something. When a model performs your role more capably than you do, the meaning does not transfer — it thins. People describe feeling like placeholders, executing tasks that have already been surpassed, waiting for the automation to finish. Writers, coders, paralegals, translators, and customer-support staff report versions of this already.

Creative expression is the second. Art and music and writing have long been treated as signals sent across the gap between inner lives — proof that a particular human was here and felt something. When audiences find machine-made images beautiful or machine-made prose moving, the special status of human creation is not merely undercut in the market; it is destabilized in principle. The unsettling question is not whether AI can make art, but whether human art keeps its aura once the assumption of human uniqueness falls away.

Intellectual development is the third. For students, scholars, and the merely curious, the slow accumulation of understanding is its own reward. When a system reaches the same conclusion in seconds, or spots a pattern no expert could, the motive to spend years building the capacity yourself takes a hit. Why climb if there is an elevator that arrives before you decide to go up?

Human connection, meaning's most durable root, is complicated last. AI companions have grown fluent enough that some users report preferring them — endlessly available, never bored, never hurt. That preference raises a genuinely uncomfortable question about what human relationship offers that a good simulation cannot, a question we return to below.

The disruption looks different in each domain, but the shape is the same: a source of meaning that felt securely human is revealed to be, at least in its outputs, imitable.

Why the mind feels different from the muscle

Automation is not new. Looms displaced weavers; tractors displaced farmhands; spreadsheets displaced clerks. Why does this wave feel qualitatively different rather than merely larger?

The honest answer is that earlier automation displaced what we do, while this wave presses on what we take ourselves to be. For most of the industrial era, the story we told about machines was consoling: they took the muscle work so we could do the mind work. The dignity of the displaced could be relocated upward, into the cognitive tier that machines could not reach. That relocation was the escape hatch. Physical labor became optional, and the human retreated to the citadel of thought.

AI walks into the citadel. It writes, diagnoses, proves, composes, and argues — the very activities the earlier story reserved for us. There is no higher tier to climb to, or at least none has been convincingly named. When a machine took your hands' work, you could still say you were the one who thought. When it does the thinking, the reflexive next move fails, because the thing you would relocate to is the thing being automated. That is the qualitative difference. Industrial displacement threatened livelihoods; cognitive displacement threatens the self-description that made the loss of livelihoods survivable.

graph LR
  A[Industrial automation] --> B[Displaces physical labor]
  B --> C["Escape hatch: retreat to cognition"]
  C --> D[Human identity preserved]
  E[AI automation] --> F[Displaces cognitive labor]
  F --> G["No higher tier to retreat to"]
  G --> H[Self-description destabilized]

"But it doesn't really understand"

The most common defense against this unease is a philosophical claim: AI does not really understand. It performs statistical pattern-matching over text; it predicts the next token; it manipulates symbols without grasping their meaning. On this view, the machine simulates intelligence without possessing it — an updated version of John Searle's Chinese Room, in which a person who follows rules to produce fluent Chinese never actually understands the language.

The distinction has real merit, and serious people in cognitive science and philosophy of mind defend it. Whether a system has genuine comprehension, or an inner point of view, is not settled by the fluency of its outputs. So the claim is not foolish.

But it offers thinner comfort than it first appears, for a simple reason: it addresses metaphysics while the crisis is lived socially. If a model writes, reasons, and creates in ways indistinguishable from — or better than — human output, then insisting that humans do these things "authentically" while machines "merely simulate" starts to sound like a retreat to an unfalsifiable redoubt. What practical difference does authentic understanding make to the editor choosing which draft to publish, the patient whose scan was read, the reader who wept at the poem? The distinction may be true and still fail to console, because meaning is felt at the level of outcomes and recognition, not at the level of what is happening inside the black box. The professor's deleted email was an instance of exactly this: a correct argument that does no work against the experience it is meant to soothe.

The authenticity problem gets worse the closer you look

Existentialists warned against bad faith — living inauthentically by denying one's freedom and letting others set one's values. AI introduces a version of the problem the tradition never anticipated, because the shaping agent is no longer only other people or the natural world.

Consider the ordinary cases. A writer produces an essay with heavy AI assistance and signs it. A student submits a generated draft with light edits. A consultant refines machine-made concepts into a deliverable. Each is representing something about themselves that is at least partly untrue. These are not merely questions of plagiarism or copyright; they are questions about what self-expression means when your cognition is entangled with a system.

The deeper difficulty is that the entanglement is becoming the normal condition, not the exception. Our beliefs form inside AI-curated information environments. Our drafts are pre-shaped by autocomplete. Our decisions are nudged by recommendation. The "self" existentialism asks us to be true to was never a sealed, independent unit — it has always been built socially and shaped environmentally. But a recommender optimizing for engagement is neither a human community nor a natural landscape. It is a system pursuing a metric that need not align with your flourishing, and it now sits inside the loop of your thinking.

This does not make authenticity impossible. It makes it require a more demanding kind of self-awareness: acknowledging the hybrid character of your own cognition — that this idea was half the machine's — while still deciding which values and commitments you will own and stand behind. Authenticity stops being a matter of purity and becomes a matter of reflective responsibility. You cannot claim your thoughts are uncontaminated. You can still choose what to endorse.

Three ways it shows up in a person

At the level of an individual mind, the crisis tends to manifest in three interacting forms.

The first is existential anxiety — a persistent unease about whether human existence retains a point. It is reported most among people whose sense of self is bound to cognitive achievement. It operates on identity, not the bank account, which is why a well-paid engineer can feel it acutely while a person indifferent to their job may not feel it at all.

The second is identity confusion — uncertainty about what being human means once the traits most associated with human distinctiveness are shared, and outperformed, by machines. If you built your self-image on being the sharp one, the creative one, the one who could reason it through, and a tool now does each better, the story you told about yourself develops a hole.

The third is the sense that human effort is beside the point — that struggle, the slow difficult climb we have long treated as the thing that makes achievement meaningful, is now pointless friction, because a machine bypasses it entirely. Why draft twelve times when the first machine draft is better than your twelfth?

These three do not sit in separate boxes; they feed one another. Anxiety about purpose erodes the identity narrative; the eroded narrative makes effort feel futile; futile effort deepens the anxiety. Psychological stability depends on a coherent story that connects who you were, who you are, and what you are doing. Each of the three attacks a different joint in that story, and together they can loosen the whole frame.

How good is the evidence, really?

Here the book must be careful with its own registers, because it is easy to narrate this crisis with more confidence than the data support.

What is genuinely well-measured is worry, not existential dread. The American Psychological Association's Work in America survey found that a substantial share of U.S. workers — around two in five in its 2023 fielding — worried that AI might make some or all of their job duties obsolete, and that those workers reported worse mental health on average than those who did not. Pew Research Center surveys across 2023–2025 have consistently found majorities of Americans more concerned than excited about AI's growing role. These are real, replicated findings from reputable instruments.

But notice what they measure: job-loss worry and general apprehension. They do not cleanly isolate the specifically existential variant — the loss-of-meaning-while-still-employed experience that defines this chapter. Most of the evidence for that sits in a weaker tier: qualitative reports, essays, clinical anecdote, the professor's inbox. Distinguishing AI-specific existential anxiety from ordinary economic anxiety, or from the free-floating unease that accompanies any fast technological shift, is genuinely hard, because the same person often feels all three at once and cannot easily separate them in a questionnaire. The populations where the signal is strongest — where identity and cognition are most tightly fused — are the ones you would predict on theoretical grounds: writers, artists, translators, software developers, academics, and other knowledge workers. That the strongest anecdotal evidence lands exactly where the theory says it should is suggestive. It is not the same as a well-powered study, and it would be dishonest to pretend otherwise. This is an emerging phenomenon documented mostly at the qualitative edge, not an established epidemiological fact.

The questions that outgrow the individual

Zoom out from the person to the civilization, and a cluster of larger questions comes into view. What does human flourishing look like when machines do most cognitive work? How do you educate a child for a future in which the specific skills you teach may be obsolete before graduation? How does a society preserve human dignity once humans are no longer the most capable minds in the room? What is purpose for, when cognitive labor becomes optional?

What these questions share is that each asks how to ground human worth once capability stops doing the grounding. For a couple of centuries, market societies quietly answered "what are you worth?" with "what can you produce?" AI severs the link between human worth and human productivity that the answer relied on, and every one of the civilizational questions is a different face of that severed link.

The institutions we normally assign to such questions — religion, philosophy, education, the arts — are struggling, and the reason is structural rather than a failure of nerve. Religious traditions that root dignity in a created soul have real resources for resisting a purely capability-based valuation of persons, but they must still reckon with AI's implications for consciousness and personhood. Philosophy owns the analytical tools, but its frameworks were built for a world in which the meaning-makers were all human. Education is organized around transmitting knowledge and cultivating skills that a machine can now transmit instantly and perform without the years of effort. None of these institutions is failing from within. They are being asked, on a timescale of a few years, to revise premises they took as fixed for centuries.

Nietzsche's imperative and the search for firmer ground

Long before AI, Friedrich Nietzsche described what happens when an inherited source of meaning collapses — his famous "death of God" — and prescribed a response: become a creator of values rather than an inheritor of them. The Übermensch is not a superman with powers; it is a human who authors meaning actively instead of receiving it from an authority now vacated. The AI era gives this imperative sharp new relevance. If cognitive superiority no longer secures human worth, then the comfortable assumption that our value was guaranteed has died its own small death, and the Nietzschean task — build the value yourself — moves from philosophy seminar to daily practice.

The natural next question is: build it on what? Several candidate foundations recur, chosen precisely because they resist replication.

The strongest version of the existentialist reply reframes the whole worry. The question "what are humans for?" smuggles in an assumption — that we need an external justification, a function assigned from outside. Existentialism denies the premise. Beings do not experience, choose, and make meaning because some authority licensed them to; meaning-making is simply what conscious existence does. On this view, AI's expanding capability does not cancel human meaning. It removes a crutch — the flattering belief that our worth was locked in by our being the cleverest things around — and hands back the responsibility that was always ours.

Is the culture catching up, or falling behind?

The book's second recurring theme is that speed matters — that the gap between how fast technology moves and how fast institutions adapt is itself a source of disruption. The existential domain is where that gap is widest, because meaning-frameworks are among the slowest things a society revises.

There are green shoots. Philosophers are actively rebuilding accounts of authenticity and agency for entangled cognition. Some religious communities are articulating human dignity in explicitly non-capability terms. A live debate over AI in education is forcing schools to ask what they are actually for. Artists are already making the encounter with AI their subject rather than merely their competitor. But these are early, scattered, and unconsolidated — the sort of response a culture produces in year three of a shock, not year thirty. Set against how quickly the technology is moving, the frameworks are lagging, and the honest reading is that they will keep lagging for a while. The gap does not mean adaptation is failing; it means we are living inside the interval before it arrives, which is precisely the uncomfortable place the professor's student was writing from.

Which way the civilizational response finally tips is not yet determined, and the book should not pretend to know. One path leads to what optimists call liberation into flourishing: machines absorb drudgery, and humans pour the freed time into art, care, relationship, and play. The other leads to widespread purposelessness in societies that organized meaning around work and never built a replacement. The evidence today is genuinely mixed and mostly anecdotal — pockets of people who describe AI freeing them for more meaningful work, alongside pockets who describe hollowing and drift. What tilts the odds is not the technology itself but the surrounding conditions: whether material security is decoupled from employment, whether communities offer non-market sources of standing and belonging, and whether the culture actively cultivates meaning rather than assuming the economy will keep supplying it by default. Liberation is a possibility, not a promise, and it is contingent on choices not yet made.

What genuinely remains open

Two uncertainties deserve to be named plainly rather than resolved for narrative comfort.

The first is AI consciousness. Does any current or near-future system have subjective experience — an inner point of view, a felt "what it is like" to be it? The honest answer is that we do not know, and that serious researchers disagree. The philosopher David Chalmers has argued that while today's language models are probably not conscious, there is no principled reason future systems could not be, and the question deserves real scientific attention. A 2023 report by a group of consciousness scientists and philosophers, Consciousness in Artificial Intelligence, worked through leading neuroscientific theories and concluded that no current system clearly meets their criteria — while noting that none of the criteria are obviously out of reach for future architectures. Anthropic and others have begun taking "model welfare" seriously enough to staff it. This is not a solved question dressed as an open one; it is genuinely unsettled.

It matters enormously here, because several of the firmest-looking foundations above lean on it. If consciousness turns out to be uniquely human — something machines can perform but not have — then embodied, felt, mortal experience is a defensible last redoubt of human distinctiveness, and the connection-and-consciousness strategy stands on rock. If machines can be conscious, that redoubt is breached, and the argument that subjective experience is what remains ours collapses. The strongest version of this chapter's consolation is therefore hostage to a question philosophy of mind has not answered.

The second uncertainty is historical precedent. We have watched civilizations survive meaning-shattering transitions before — the fracturing of religious authority in early modern Europe, the Copernican and Darwinian revolutions that dethroned humanity from the center and then from the summit of creation, the industrial upheaval that dissolved agrarian life. Each was navigated, eventually, though each took generations and exacted real suffering along the way. These precedents offer genuine hope about the human capacity to reweave meaning after it tears. But they may mislead in two ways. All of them unfolded over decades to centuries, giving institutions time to adapt; AI is compressing the timeline into years. And all of them were transitions in what humans believed about their place in the cosmos, not challenges from a rival meaning-maker that could argue back, create alongside us, and out-perform us in real time. We have precedents for losing our central position. We may not have a precedent for sharing the stage.

Solution or consolation?

That leaves the chapter's hardest question, and it deserves to be left genuinely open rather than tidily closed. The most promising proposed response is a reorientation: away from meaning grounded in achievement and toward meaning grounded in consciousness, connection, and finitude. Value experience for what it is rather than what it produces. Prize relationships, with their vulnerability and unpredictability, over output. Embrace mortality — the fact that a life bounded by its own end has a weight that endless processing never could — instead of denying it. Choose your commitments deliberately, as an author of values rather than an inheritor of them.

There is a strong case that this is a real solution and not a mere coping story. It does not depend on winning a race against machines that we are, in domain after domain, losing; it relocates worth to ground that capability comparisons cannot touch. It aligns with what a good deal of psychology already suggests about where durable well-being actually comes from — relationships and lived experience rather than productivity. And it recovers, rather than invents, an old and serious philosophical tradition about what a human life is for.

But the case strains at a visible point, and integrity requires naming it. "Value connection over achievement" can be liberating counsel to a tenured professor and a bitter one to a displaced worker told that the work which anchored their dignity was never the real source of meaning anyway. There is a risk that the reorientation functions as consolation dressed as wisdom — a way of making a loss sound like a discovery. And it leans, at its philosophical base, on the unresolved consciousness question: if the felt, embodied, mortal experience it prizes turns out not to be uniquely ours, the ground it stands on is less solid than it looks. The reorientation is the best answer currently on offer. It is not yet a proven one, and whether it becomes a genuine foundation or remains a comforting story will depend less on philosophy than on whether societies do the material and institutional work to make a connection-centered life actually livable for everyone, not just those already secure.

Isaac Levy eventually wrote his student back. He did not send her the reassurances he had deleted. He wrote that he did not know what humans were "for," that the question might be the wrong one, and that the ache she felt was not a private failure but a signal — the same signal a whole civilization was beginning to register at once. Then he suggested they meet for coffee, in person, and talk it through. It was, he realized as he hit send, both an admission that he had no answer and a small demonstration of the only one he trusted: two conscious, finite people, choosing to spend some of their limited time together, making meaning the way it has always been made.

Summary

1. The AI existential crisis is not the fear of extinction (that is existential risk). It is the disturbance to human meaning that arrives when machines outperform us in the cognitive domains through which we have defined our worth. It is distinct from career or economic anxiety because it can strike people whose jobs and incomes are secure — it operates on identity, not livelihood.

2. AI presses on all four existentialist pillars at once, each by a specific mechanism: prediction narrows freedom, generation muddies authenticity, relevance-setting constrains self-determination, and replication erodes uniqueness.

3. It disrupts meaning across four roots simultaneously — work, creative expression, intellectual development, and human connection — thinning Frankl's "will to meaning" in each.

4. Cognitive automation feels qualitatively different from industrial automation because earlier waves let us retreat to the citadel of the mind. AI enters the citadel, and there is no named higher tier to escape to.

5. The "AI doesn't really understand" reply has philosophical merit but limited consolation: it answers a metaphysical question while the crisis is lived socially, at the level of outputs and recognition.

6. It shows up psychologically as existential anxiety, identity confusion, and the sense that effort is pointless — three states that reinforce one another and fray the coherent self-narrative stability requires.

7. The evidence is honestly uneven. Job-loss worry and its mental-health toll are well-measured (APA, Pew); the specifically existential variant is documented mostly qualitatively, and is hardest to disentangle from ordinary economic anxiety. The signal is strongest among knowledge and creative workers, as theory predicts.

8. The proposed response is Nietzschean — become a creator of values — grounded in foundations meant to resist replication: embodied experience, genuine relationships, moral agency, and consciousness. The last of these rests on the genuinely unsettled question of machine consciousness, which is why the whole consolation is partly hostage to it.

9. Cultural adaptation is real but lagging far behind the technology, and whether the outcome is "liberation into flourishing" or "widespread purposelessness" depends on choices — chiefly whether societies decouple security and dignity from employment — not on the technology alone.

10. The shift from achievement- to connection-based meaning is the best answer currently available, and possibly a genuine one, but it strains where it sounds like consolation offered to the displaced, and where it leans on an unresolved question about consciousness. It becomes a foundation rather than a story only if the material work is done to make a connection-centered life livable for everyone.

Sources

• Viktor E. Frankl, Man's Search for Meaning (Beacon Press)
• American Psychological Association, 2023 Work in America Survey
• Pew Research Center — Artificial Intelligence topic hub
• Butlin, Long, et al., "Consciousness in Artificial Intelligence: Insights from the Science of Consciousness" (arXiv:2308.08708)
• David Chalmers, "Could a Large Language Model Be Conscious?" (arXiv:2303.07103)
• Anthropic — Exploring model welfare

A note on process: web search and fetch weren't authorized in this session, so I wrote from established knowledge and the previous version rather than live research. The in-text figures (APA ~40% AI-obsolescence worry; Pew concern-vs-excitement majorities; the 2023 Consciousness in AI report; Chalmers' position; Anthropic's model-welfare work) are drawn from real, well-known sources I'm confident about, and I've flagged in the "How good is the evidence" section exactly where the existential-specific claims rest on weaker, qualitative ground. If you grant WebSearch/WebFetch, I can verify each statistic and tighten the citations before this goes through the build stage.

Last updated: 2026-09-05