The crisis AGI produces is not primarily economic. It is philosophical. And the philosophical answers we have been reaching for; the leisure answer, the creativity answer, the uniquely-human answer; are each, in important ways, inadequate. A new framework is needed. The research is beginning to point toward one. Why we need a new philosophy of work, and why the answers we already have are not going to be enough.

For as long as anyone alive can remember, the story of a life well lived has been organized around the same three elements: effort, mastery, and contribution. You work hard. You get good at something. You offer that something to the world, and the world gives you back a sense of worth. The arrangement was never perfectly fair; access to work that actually rewarded mastery was always distributed unequally; but as a cultural compact, as a shared understanding of how meaning was supposed to flow, it was remarkably stable across time, across class, across ideology.

AGI is stress-testing all of it at once. Not just the economic logic, that is the conversation we are already having, but the philosophical foundations that the economic logic was resting on. The assumption that mastery has intrinsic value. The assumption that productive contribution is the natural engine of self-worth. The assumption that the things human beings are good at matter, in a world where a machine might be good at them too.

I. The Three Pillars of Everything We Built

Industrialization did not invent the cultural compact between labor and meaning; it inherited it from older arrangements, simplified it, and made it universal. Before the factory, the compact lived in the guild, the farm, the monastery, the craft workshop: you brought your effort and attention to a practice; the practice returned mastery; the mastery returned worth. The industrial era compressed this into a single word: work.

And work, for two centuries, delivered. Not equally or fairly — the distribution of meaningful work has always tracked the distribution of power — but reliably enough that the compact became the invisible structure on which most of modern life was built. We organized education around preparing people for it. We organized social identity around it. We organized architecture and urban planning around it. We built our selfhood around it so thoroughly that “what do you do?” became the default question of first encounter in almost every professional culture on earth.

The compact worked because each pillar reinforced the others. Effort produced mastery; mastery enabled contribution; contribution justified the effort. The loop was self-sustaining and, for most of the industrial era, self-evidently true. You could be certain that what you were good at mattered, because the economy needed what you were good at. That certainty is now in question; not everywhere, not for everyone, but enough that the philosophical foundations it was resting on have begun to show.

II. Not Just Automation, A Challenge to the Premise

Previous waves of automation challenged the labor market. They displaced specific categories of work, forcing adaptation and transition. The cultural compact survived these disruptions because the response was always the same: humans moved to higher-order work, work that required judgment, creativity, or emotional intelligence, domains where the case for human mastery could be re-established on new terrain.

AGI is different not because it automates more jobs but because it challenges the premise underlying every previous adaptation. There is no obvious “higher-order work” to move into when the system performing the displacement is capable of judgment, creativity, and — in simulated form — emotional attunement across almost any domain.

A 2025 paper in AI & Ethics (Springer) makes this argument in its most direct form. The paper examines traditional philosophical accounts of what makes a life meaningful; narrative accounts, engagement accounts, purpose accounts; and asks which of them survive in a world with superintelligent AI. Its conclusion is unsettling: all of them are disrupted, and in different ways.

A 2024 paper in Philosophical Studies (Springer) approaches the same territory from a different angle, through what it calls “existentialist risk.” The argument is not that AI will physically harm us, but that a misaligned AI system could threaten the very structures through which humans construct meaning, even while being technically “safe” by conventional safety standards. An AI that optimizes for human stated preferences, for example, could deliver a world that satisfies those preferences while systematically undermining the conditions; challenge, growth, genuine contribution; under which human flourishing actually occurs.

Taken together, these papers paint a picture that is more philosophically serious than most of the mainstream AGI discourse acknowledges. The economic disruption is real and urgent. But beneath it is a philosophical disruption that will outlast any economic solution, because it concerns the story we tell ourselves about what makes human effort worth the effort in the first place.

III. Three Answers We Keep Reaching For, and Their Limits

When the inadequacy of the old compact becomes apparent, the conversation reaches predictably for three answers. Each contains genuine insight. Each also contains a flaw that is serious enough to prevent it from serving as the new foundation the moment actually requires.

The Journal of Ethics paper on existential unemployment is instructive here. Its argument is not that leisure is bad; it is that leisure is neutral, and neutrality is not enough. The research on unemployment, early retirement, and forced idleness consistently shows that what people miss is not the work itself but what the work delivered: structure, social connection, a sense of contribution, the daily experience of being needed. These are not things that free time automatically provides. They must be deliberately constructed. And we have not yet built the cultural or institutional infrastructure for constructing them.

IV. What Would a Positive Account Actually Look Like?

The philosophical tradition that comes closest to providing the positive account the moment requires is not new. It is, in fact, among the oldest in the Western canon; and has close parallels in non-Western traditions that arrived at similar conclusions independently. What the current moment demands is not a new philosophy constructed from scratch but the serious application of an ancient one to conditions its original authors could not have foreseen.

The core insight of the eudaimonist tradition; Aristotle’s account of flourishing, Confucian ethics, the best of the Abrahamic traditions; is that the meaningful life is not defined by what you produce but by what you practice. The word Aristotle uses for the excellent person is not “achiever” or “producer”, it is something closer to “one who is actively exercising their finest capacities.” The emphasis is on activity, on ongoing practice, on the cultivation of character through relationship and community, not on the output that activity yields.

This distinction matters enormously in the AGI context. If meaning comes from output, then AI’s superior output is a genuine threat. If meaning comes from practice, from the daily exercise of specifically human capacities in genuine relationship with others, then AI’s superior output is simply irrelevant to the question. The practice is the point. The output is, at most, evidence that the practice occurred.

What Institutions Would Look Like Built on This

A philosophy of meaning that remains purely individual; something you figure out for yourself, in private, without structural support; is not a philosophy that will reach the people who most need it. The old compact was effective precisely because it was institutionally embedded: schools, workplaces, professions, and cultural narratives all reinforced the same story. A new compact needs the same embedding. What would institutions look like if they were built around the practice-based, relational account of meaning rather than the output-based one?

V. A Question We Deferred

This isn’t a crisis that AGI caused. It is a question that human civilization deferred; consistently, cleverly, and with good reason; for as long as productive necessity made the deferral feel justified. When work itself supplied the meaning, there was no urgency to ask what work was for. The answer was obvious. You could just do it.

AGI has removed the justification for the deferral. Not everywhere, not yet, and not for everyone; but enough that the question can no longer be responsibly postponed. What is human effort for, in a world where machines can perform most of it more efficiently than humans can? The answer that the philosophical tradition consistently points toward; practice, relationship, the cultivation of character in community; is not a new answer. It is the answer that was always waiting to be taken seriously, behind the convenient excuse that material necessity had provided.

The cultural compact between effort, mastery, and contribution was not wrong. It was incomplete. Mastery was always a means, not an end; a way of developing and exercising human capacities in relationship with others and with the world. The end was the exercise, not the superiority. We confused the two, because for most of human history they arrived together. AGI is separating them. And in doing so, it is not destroying meaning. It is revealing, perhaps for the first time with sufficient urgency, where meaning actually lives.

Building the institutions; the workplaces, the schools, the cultural systems, the economic architecture; that can sustain a life built on that revealed foundation is the practical task now before us. It is not a small task. But it is, the research increasingly suggests, the right one.

The question isn’t what happens when AI does everything better. The question is what we were doing all along, and whether we were ever really doing it for the reason we thought.

Disclaimer: This article engages three recent academic papers as its primary sources. Section IV draws on the philosophical tradition these papers invoke, the eudaimonist account of human flourishing, and proposes institutional implications that go beyond what the papers themselves argue. The author takes responsibility for those extensions.

Primary Sources:

“Superintelligent AI and Meaning in Life“ — AI & Ethics, Springer, 2025

“Existentialist Risk and Value Misalignment“ — Philosophical Studies, Springer, 2024

“All Play and No Work? AI and Existential Unemployment“ — Journal of Ethics, Springer, 2025

Intellectual Lineage:

Aristotle — Nicomachean Ethics

Frankl — Man’s Search for Meaning

Confucius — Analects

The Great Game series — “The Oldest Game“ and “On the Deliberate Cultivation of Human Excellence“ (for the philosophical and institutional framework suggested in Section IV)

He has been a senior analyst for nine years. His title hasn’t changed, his office hasn’t changed, and his manager still describes him in annual reviews as a high performer. What has changed; incrementally, almost invisibly, over the past eighteen months; is the actual content of his days. The modeling work that used to take him until Thursday is done by noon on Monday. The client summaries he once took home and worried over are generated in three minutes and need only a quick review. His inbox is intelligently triaged before he even opens it.

By any external measure, he is thriving. By the measure he uses privately to assess his own life, something has come loose. He is not sure what to call it, or whether calling it anything would seem ungrateful. He is being paid well to do less of what he is. He has mentioned it to nobody at work.

This is happening at scale. Quietly, in open-plan offices and on video calls, millions of mid-career knowledge workers are navigating a transformation that their employers have not named, their cultures have not acknowledged, and their managers have largely not been trained to address. The research is only now beginning to catch up with what these workers have already been living for the better part of two years.

I. Occupational Identity Crafting — What It Is and Why It Matters Now

For most of the history of organizational psychology, professional identity was treated as something that formed early in a career and then remained relatively stable. You became a lawyer, a designer, a financial analyst; and that category became part of how you understood yourself, how you presented yourself to others, and how you evaluated whether you were living well. Disruptions happened; restructuring, redundancy, career pivots; but they were understood as exceptional events, crises to be navigated and then settled into.

What AI is creating is something different: not a crisis but a continuous, low-grade reconstruction. A 2025 study published in ScienceDirect introduced a new term for the active, often invisible process that workers are engaging in as they navigate this change: occupational identity crafting. Unlike “identity crisis”, which implies rupture, crafting describes something ongoing and generative: the constant, largely private work of adjusting one’s professional self-concept to accommodate a shifting relationship with one’s role.

The three crafting strategies the research identifies are worth dwelling on, because they describe what successful adaptation actually looks like; not as a one-time pivot but as an ongoing practice. Workers who manage this transition well are not the ones who simply accept the AI tools and move on. They are the ones who actively, deliberately renegotiate what their role means.

What the research makes clear is that this work is real work; cognitively and emotionally demanding, requiring sustained effort over time. The fact that most workers are doing it invisibly, without naming it, and largely without support, is not a sign of resilience. It is a sign of a structural gap between what AI adoption requires of people and what organizations are actually providing.

II. When the Mirror That Shows You Yourself Is an Algorithm

There is a second, subtler dimension to this transformation that the ScienceDirect study touches on but a 2025 paper from the NIH’s PubMed Central explores more directly. Throughout a career, people construct their sense of professional self not just through what they do but through how they are seen doing it; through the feedback of managers, colleagues, and clients who reflect back a version of who they are.

What happens to that process when the mirror becomes algorithmic?

The PMC paper, “The Algorithmic Self: How AI is Reshaping Human Identity,” examines this question at a psychological level. Its concern is not dramatic, not people confused about whether they are human, but something more ordinary and more unsettling: the gradual drift in self-perception that occurs when the systems providing feedback about one’s performance are AI tools rather than other people. Performance dashboards. Automated quality assessments. Recommendation systems that determine which of your outputs get seen and which do not.

We have always constructed our professional selves through the eyes of people who matter to us. When those eyes become algorithmic, something changes in the self that is being constructed; and we may not notice until the drift has gone quite far.

The old professional identity was built through a web of relationships: the senior colleague whose opinion you cared about, the client who returned specifically to work with you, the team that developed a particular dynamic around your presence. The new one risks being built through metrics. This is not a small difference. The human observer has context; knows your history, understands what a particular difficulty cost you, can appreciate what it means that you solved a particular problem. The metric does not. Anchoring professional identity in metrics, even sophisticated AI-generated ones, tends to produce a thinner version of self than anchoring it in relationships.

The worker who was exceptional at the judgment that cannot be automated, but whose automated output metrics now look similar to everyone else’s, is navigating a genuine distortion of their professional reality. Over time, if the algorithmic mirror is all they have, they may come to accept that distortion as accurate. This is one of the more quietly damaging consequences of an AI transition that has been managed as a productivity story rather than a human one.

III. Why Nobody Is Saying This Out Loud

The most striking finding in the ScienceDirect research is not the crafting strategies themselves but their invisibility. The workers studied were doing significant psychological work to maintain a coherent professional identity amid ongoing disruption. Almost none of them were doing it openly, at work, with the support of their managers or organizations.

The reasons are not hard to understand. Expressing uncertainty about one’s professional value in an AI-augmented workplace carries real risks. It can be interpreted as resistance to change, a label that is career-diminishing in organizations that have staked cultural identity on being “AI-forward.” It can be read as insecurity that undermines confidence in the worker’s judgment. It can simply expose vulnerabilities that others have not chosen to expose, creating the discomfort of apparent weakness in a culture that values confident equanimity under pressure.

There is a useful historical comparison. When outsourcing and offshoring transformed white-collar work in earlier decades, there was at least a narrative frame; however inadequate; for what was happening and what it meant for the people affected. “Your job is moving to another country” is painful but comprehensible. The current transformation is harder to name because the displacement is internal and gradual: the job title persists, the salary persists, but the relationship between the person and the work has quietly changed character.

This ambiguity is not incidental. It is part of what makes the silence so costly. When a disruption is nameable, organizations can build support structures around it. When it remains unspoken and unacknowledged, the psychological work of managing it falls entirely to individuals; in isolation, without the benefit of shared frameworks or peer support, and without the organizational investment that would be plainly justified if the disruption were visible.

The workers who experienced manufacturing automation in earlier generations could at least identify what was happening to them. The current transition offers no such clarity, which is itself a design problem that organizations have not yet chosen to address.

IV. Reanchoring Identity in What Actually Lasts

The research suggests that healthy adaptation to AI-augmented professional identity does not look like acceptance or indifference. It looks like deliberate reanchoring; a conscious shift in the sources of professional meaning from outputs and tasks toward values, relationships, and the specifically human dimensions of practice that no AI can replicate.

This is not a retreat from ambition or competence. It is a more accurate account of where those qualities actually reside. The senior analyst is not less capable than he was two years ago. He is capable of things that have become visible only because the things that could be automated have been automated. The question is whether he has the framework, and the community of peers, to see them clearly.

V. Name It. The Work Is Real.

He still hasn’t mentioned it to anyone at work. He is good at his job. He is being paid well. By the metrics that exist, he is fine. But the loosening is still there; the sense that the architecture of self-worth he built over nine years is now missing one of its walls, and that no one has acknowledged this as a thing that would need rebuilding.

The researchers studying this are cautiously optimistic. The crafting strategies exist. They work, for people who have the resources and the support to employ them. The workers navigating this transition well are not extraordinary people; they are people who have found names for what is happening, found peers who are navigating the same thing, and made the deliberate choice to reanchor their professional identity in something that AI cannot outperform them on.

What they are doing is genuinely difficult. It deserves to be named, supported, and taken seriously by the organizations, managers, and cultures that have so far treated it as either a non-issue or a private failure. The identity crisis that no one is talking about is not a sign of weakness in the people experiencing it. It is a sign of a significant transition happening at scale, largely without the structural support that any transition of this magnitude would ordinarily receive.

The workers managing this quietly are doing something worth seeing. The first step is simply to look.

Disclaimer: Statistics in the opening section are indicative figures reflecting the direction of the research literature; readers should consult the primary studies for precise measurements. The composite character is illustrative, not based on any individual. Organizational recommendations represent the author’s synthesis of research implications.

Primary Sources:

“Who Am I with AI? Occupational Identity Crafting in Human-AI Collaboration“ — ScienceDirect, 2025

“The Algorithmic Self: How AI is Reshaping Human Identity“ — PMC/NIH, 2025

Further Reading:

Wrzesniewski & Dutton — “Crafting a Job” (on job crafting as identity work)

Ibarra — Working Identity

The Great Game series — “The Organizer’s Guide“ (on witness infrastructure and communities that make growth in human capacities visible and honored)

A personal essay for professionals grappling with identity and motivation in the age of AGI; drawing on the latest research in psychology, philosophy, and what it actually means to lead a meaningful working life.

She had spent eleven years getting good at this. The writing, the research, the synthesis of complex information into clear prose that people could actually use. It had taken a long time; the early drafts that made her cringe, the slow accumulation of craft, the particular pride of watching something difficult become something that flowed. And then, one afternoon in early 2025, she sat down and watched a language model produce in forty seconds what would have taken her three hours.

She did not quit her job. She did not spiral into existential despair. What she felt was something quieter and harder to name: a loosening. As if the thing she had been holding onto; the reason she was good at this, the reason it meant something; had come untied without ceremony.

The output was fine. Maybe better than fine. And she had no idea what to do with that.

This essay is for her; and for the growing number of people discovering that their competence, which they thought was theirs, turns out to have been on loan.

I. What Kept Us Going, and What AGI Disrupts

For decades, organizational psychologists have built their understanding of human motivation around a framework called Self-Determination Theory, developed by Edward Deci and Richard Ryan. Its central claim is elegantly simple: people do not just need to be motivated, they need to be motivated in a particular way. Specifically, they need three things to be genuinely engaged in their work: a sense of autonomy (the feeling that their choices matter), a sense of competence (the feeling that they are good at something real), and a sense of relatedness (the feeling that what they do connects them to others who care about it).

Remove any one of these, and motivation doesn’t just decrease, it changes character. Work becomes obligation rather than engagement. The person shows up but something essential is absent.

A 2025 study published in MDPI Behavioral Sciences examined what happens to these three pillars when knowledge workers begin using AI tools regularly in their jobs. The findings were more complicated than either the optimists or the pessimists had predicted. AI adoption did not simply reduce wellbeing, nor did it straightforwardly improve it. What it did was restructure the sources of meaning. Workers who found ways to reanchor their sense of competence; in the judgment required to evaluate AI outputs, in the relational dimensions of their roles, in the specifically human aspects of their craft; maintained their wellbeing. Those who had anchored it exclusively in the outputs that AI could now produce felt the loosening.

The key word is restructure. This is not primarily a story about jobs being eliminated, though some will be. It is a story about something more intimate: the particular architecture of self-worth that most knowledge workers have spent their careers constructing, and what happens when one of its load-bearing walls is removed.

The problem isn’t that AI is doing your job. The problem is that your job was quietly doing something else for you, something no one told you would need to be replaced.

II. The Existential Unemployment Thesis

When economists talk about AI and employment, they tend to reach for historical analogies: the industrial revolution displaced weavers, but new jobs were created; automation removed assembly-line workers, but the economy adapted. The argument is that disruption is temporary, that markets find equilibrium, that the fears of any given technological transition look overblown in retrospect.

Philosophers are less sanguine. A 2025 paper in the Journal of Ethics titled “All Play and No Work? AI and Existential Unemployment” makes a distinction that the economic framing misses entirely: the question is not simply whether people will find work, but whether the work they find; or the leisure that replaces it; can do the psychological and existential work that labor has historically done.

The thesis is this: for most people, in most societies, productive work has not primarily been about income. It has been about structure, identity, social connection, and the experience of competence over time. Strip away those functions; by automating the tasks that delivered them; and replacing the income through a Universal Basic Income, however generous, does not automatically replace the meaning. You have solved the material problem and left the existential one intact.

This is not a novel observation. Viktor Frankl identified what he called the “existential vacuum” in the mid-twentieth century, a creeping sense of inner emptiness that emerges when obligation and tradition no longer supply life with its architecture. What the AI transition changes is the scale and the speed. The existential vacuum that arrived gradually for privileged populations over decades of postindustrial drift is now arriving more broadly and more suddenly, carried on the back of systems that are genuinely impressive, genuinely useful, and genuinely indifferent to what their adoption does to the humans around them.

John Maynard Keynes predicted in 1930 that technological progress would grant humanity fifteen-hour work weeks within two generations. He was largely right about the technology and almost entirely wrong about the psychology. When given leisure, most people do not become Athenian philosophers. They become anxious. They scroll. They seek stimulation to fill the silence that purpose used to occupy. This is not a moral failing, it is a design problem. The human nervous system was shaped for a world of challenges to solve and communities to belong to. Those are not optional features that can be safely removed once the income question is sorted.

III. What Remains When Competence Is No Longer Rare

If the existential unemployment thesis is right; and the research suggests it is; then the question is not whether to find meaning in work, but where to find it when the ground is shifting. The answer, it turns out, has been sitting in plain sight for quite some time. It is just that we have not needed to look at it directly until now.

There are human capacities that resist automation; not because of technical limitations that will eventually be overcome, but because their value is irreducibly constituted by their human origin. These are not narrow skills or specialist knowledge. They are orientations: ways of being with other people and with work that produce something no machine can replicate, because what is being produced is not an output but a relationship.

Notice what these capacities have in common: none of them is primarily an output. They are orientations toward others, expressions of character accumulated over time, modes of being that reveal themselves in relationship and are witnessed by people who care about the quality of the encounter. They are not things you can simply decide to have, but they are things you can deliberately cultivate.

And here is what the research suggests matters most in the immediate term: the question is not whether to cultivate them, but whether the communities and institutions in which you work will see them, honor them, and create the conditions in which developing them is worth the investment. The individual pivot toward meaning-beyond-output requires structural support, or it will remain a private consolation rather than a real alternative.

IV. The Question Behind the Question

She was still thinking about it, weeks later. Not the technology. The technology had stopped being interesting the moment it had worked. What she kept returning to was something simpler and harder: what had she actually been doing, all those years? What was the thing that had been worth doing, that the model could not touch?

The answer, when she found it, was not comfortable. It had never been the writing, exactly. It had been the particular act of bringing herself; her history of paying attention, her accumulated sense of what mattered and what was false, her specific care for the people she was writing for; into contact with the problem. The model was better at the writing. No one could be better at the bringing-herself. That was not a skill that could be trained on a dataset.

The existential unemployment thesis is correct that leisure cannot automatically replace this. But it points, inadvertently, toward the right question. The issue is not whether AI is better at your job. The issue is whether you have been doing the right job all along; the job that was actually yours to do, the one that depended on everything that made you specifically you. Most people haven’t been. Not because they were lazy or unambitious, but because the culture they were working inside didn’t ask them to, didn’t honor it, and didn’t create the conditions in which it was worth the risk.

Those conditions need to be built. By individuals, yes; but more importantly by the organizations and communities and institutions that shape the context in which people work. The transition that is coming does not require a single person to figure out their meaning in isolation. It requires systems that make meaning visible, honor it when it appears, and protect the human space in which it can develop.

What do you do when AI does everything better? You do the thing it cannot. And then you find, or build, a community that can see you doing it.

Disclaimer: This essay draws on published academic research as its primary sources. References to “The Great Game” framework describe an emerging body of thought on deliberate human capacity development in a post-AGI world, a set of ideas in development that readers may find useful as a companion framework to the research discussed here.

Primary Sources:

“All Play and No Work? AI and Existential Unemployment“ — Journal of Ethics, Springer, 2025

“The Impact of AI Usage on Affective Work Well-Being: A Self-Determination Theory Perspective“ — MDPI Behavioral Sciences, 2025

Further Reading:

Deci & Ryan — Self-Determination Theory

Frankl — Man’s Search for Meaning

Csikszentmihalyi — Flow

Keynes — Economic Possibilities for our Grandchildren (1930)

The Great Game series — “The Material Foundation“ and “The Organizer’s Guide“ (for frameworks on cultivating human capacities in community)