“When you take away someone’s work, you take away their story. Their daily rhythm, their source of competence, their reason for leaving the house, their way of explaining themselves to others and to themselves. That is not a productivity question. It is an existential one. And we are not having that conversation at anywhere near the required depth.”
Neal Lloyd · Inside The Machine, Day 24In 1930, the economist John Maynard Keynes published an essay called “Economic Possibilities for Our Grandchildren.” He predicted that by 2030, advances in technology and productivity would reduce the standard working week to fifteen hours — that the economic problem, as he called it, would essentially be solved, and that humanity would face the far more difficult challenge of figuring out what to do with its leisure. Keynes was wrong about the fifteen-hour week. He was not wrong about the underlying dynamic. Technology has continuously increased productivity and continuously generated new forms of work — so far. The question that AI poses with a sharpness that no previous technology has quite matched is whether this time is different: whether AI is sufficiently general and sufficiently capable that it breaks the historical pattern in which new technology destroys some jobs but creates others, leaving the total volume of meaningful human work approximately constant. This is Day 24 of Inside The Machine. We are not going to answer the economic question today — nobody can, with confidence. We are going to address the question underneath it. What is work actually for?
The Functions of Work Beyond the Payslip
The economic function of work is well-understood and well-measured: it provides income that enables consumption, which drives economic growth. But work performs several functions that are not captured in the income measure and that the economic debate about AI-driven displacement largely ignores.
Structure and temporal organisation. Work imposes a temporal structure on the day, the week, and the year that most people do not consciously appreciate until it is removed. The rhythm of working days and rest days, the sequence of tasks that constitute a job, the meetings and deadlines and deliverables — these provide what psychologists call “time structure,” a scaffolding of meaning that orients daily life. Studies of long-term unemployment consistently find that the loss of time structure — not just the loss of income — is one of the most distressing aspects of joblessness. People without work do not simply relax with their leisure. They experience the unstructured time as disorienting, empty, and demoralising. Time without purpose is not freedom. It is a specific kind of suffering.
Social connection and collective purpose. Work is one of the primary mechanisms through which people form relationships outside their immediate families. Colleagues, professional networks, the shared project of an organisation pursuing a goal — these are sources of social connection and collective purpose that many people depend on more than they consciously realise. The shift to remote work during the 2020s revealed this in stark terms: productivity was maintained but loneliness increased, and many workers reported missing the social texture of working with others even when they did not miss the commute or the office politics. Work provides belonging as well as income.
Competence and mastery. Work provides regular opportunities to develop skill, to demonstrate capability, and to experience the satisfaction of mastery — of becoming better at something over time. The psychological literature on wellbeing consistently identifies competence as one of the three core human needs (alongside autonomy and relatedness). Work is, for most adults, the primary domain in which the competence need is systematically met — through learning, through promotion, through the feedback of doing something well and having that recognised.
Status and identity. “What do you do?” is among the first questions adults ask each other in social settings across most of the world’s cultures. The answer is not merely a description of time allocation — it is an identity statement. Occupational identity — being a nurse, an engineer, a teacher, a builder — is one of the primary ways people understand and present themselves. The loss of a job is not merely an income loss; for many people it is an identity loss of significant psychological magnitude. The research on unemployment and mental health is unambiguous: sustained joblessness is associated with significantly elevated rates of depression, anxiety, and substance abuse, at levels that persist even after controlling for income loss.
Time structure: daily rhythm, weekly pattern, year-shaped by work cycles. Social connection: colleagues, professional networks, shared purpose. Competence: skill development, mastery, recognised capability. Status: “What do you do?” — occupational identity as primary social self-presentation. Purpose: contribution to something beyond oneself, being needed. Health: long-term unemployment is independently associated with elevated mortality, depression, and substance abuse after controlling for income. The income function of work is what policy debates focus on. The identity, structure, social, and purpose functions are what people mourn when work is taken away. They are also what no Universal Basic Income payment directly replaces.
The Competence Question in an Era of Machine Superiority
One of the most psychologically significant features of the current AI transition is the nature of the capabilities that are being automated. Previous waves of automation largely targeted physical and routine cognitive tasks — assembly line work, data entry, customer service scripts. These tasks provided income but were not typically sources of high occupational pride or identity. People who lost these jobs experienced real hardship but the jobs themselves were not core to most workers’ sense of professional identity in the way that knowledge work is.
The AI transition is targeting knowledge work directly — the tasks that lawyers, doctors, analysts, writers, engineers, and researchers identified as the core of what they did and the basis of their professional identity. When AI can write briefs, generate analyses, produce code, and draft medical notes faster and often better than the human professionals who trained for years to perform these tasks, something psychologically significant happens beyond the employment economics. The foundation on which those professionals built their sense of competence and value is eroded. The skill they spent a decade developing is revealed to be more substitutable than they believed. The identity they built around professional expertise is suddenly contested in a way it was not before.
This is not a hypothetical. The entry-level hollowing that Ground Truth’s Episode 05 documented — the 13-20% employment decline for young workers in AI-exposed occupations — means that people at the beginning of their knowledge work careers are being denied the developmental experiences through which professional identity is built. The junior analyst who never gets to do junior analyst work cannot develop the skills, the judgment, and the professional identity that the senior analyst role eventually requires. The pipeline is being blocked at exactly the moment when the people entering it most need the developmental experiences it provides.
The jobs that previous automation targeted were sources of income. The jobs that AI is targeting are sources of identity. The lawyer who loses conveyancing work to AI has lost more than a revenue stream — they have lost a component of what it meant to be a lawyer. The analyst who loses research tasks to AI has lost more than time — they have lost the developmental experience through which analysis judgment is built. This is a different kind of displacement than the one we have managed before.Neal Lloyd · Inside The Machine, Day 24
The Industrial Revolution Comparison and Why It Both Helps and Misleads
The standard reassurance in the AI jobs debate is to invoke the Industrial Revolution: automation destroyed agricultural and craft jobs and created factory and service jobs; the total volume of employment increased and average living standards improved dramatically over a century. The same pattern, the argument runs, will repeat with AI — jobs destroyed in one domain will be offset by jobs created in new domains we cannot yet fully anticipate, and the net effect over the long run will be positive.
The historical analogy is real and provides genuine grounds for cautious optimism about the very long run. It is also limited in several important ways. First, the timescale: the transition from agricultural to industrial employment took approximately a century. The economic history of that century includes the Irish Famine, repeated financial panics, child labour on an industrial scale, urban poverty that was worse than rural poverty for a generation, and political violence across the European continent. The long run was positive. The transition was brutal for the people living through it, and “technology always creates more jobs than it destroys over the long run” was cold comfort to the handloom weaver whose livelihood was destroyed in the 1820s and who did not live to see the 1920s.
Second, the nature of the transition: the Industrial Revolution created new forms of work that were different from what it destroyed but were similarly accessible to the people it displaced. The agricultural worker could, with difficulty and adjustment, become a factory worker. The factory worker’s children could become clerks and tradespeople. The transitions were painful but the new work was not categorically more cognitively demanding than what it replaced. The AI transition is targeting cognitive work — the domain that all previous transitions produced as the safe harbour for humans displaced from lower-skill roles. If AI is sufficiently general, there may not be an analogous safe harbour to move to. The new jobs may require cognitive capabilities that the displaced workers do not have and cannot easily develop. The analogy breaks down at exactly the point where it is most needed as reassurance.
Third, the speed: a century-long transition, however brutal, allowed institutions — educational systems, labour markets, social support structures, cultural norms about work — to adapt over generations. A transition that plays out in a decade does not allow this. The institutions are not failing to adapt to AI because the people running them are incompetent. They are failing because the speed of the change is categorically faster than the speed at which institutions designed for the human timescale can operate.
Beyond Jobs Policy: The Questions That Actually Need Answering
The income question and the meaning question are not the same question. The policy debate about AI and work is dominated by the income question: how do we ensure that people displaced by AI have adequate economic resources? UBI, expanded social insurance, retraining programmes, profit-sharing from AI companies — these are all income-focused responses to an income-focused framing of the problem. They are necessary. They are not sufficient. The identity, structure, social connection, and competence functions of work are not provided by a cash payment. They require different policy thinking — about how people form communities of shared purpose, how meaningful contribution is recognised and valued, and how the developmental experiences that work provides can be maintained even as the income-generating activities shift.
The care and relational sector is systematically undervalued and underemployed. The work that AI is worst at — emotional presence, genuine human connection, situated judgment in complex interpersonal situations — is the work that the economy currently values least. Care workers, teachers, community organisers, mental health workers — these roles provide some of the most important functions that work performs, for the workers performing them and for the people they serve, and they are compensated at a fraction of the rate of the knowledge work that AI is displacing. A rational response to AI displacement might involve dramatically re-valuing the care economy — not as a second-order consolation prize for people who cannot compete with AI but as a primary sector that produces things AI cannot and that the economy currently produces less of than it should. This would require political and cultural change that is more difficult than any technical problem.
The developmental question requires answers before the displacement arrives. The most urgent practical question is not what to do when people have been displaced but how to ensure that people currently in the workforce — and people entering it — are developing the capabilities that remain distinctively valuable as AI capabilities expand. Deliberate friction — the practice of maintaining and developing human skills even in domains where AI can perform the same task faster — is the concept that several researchers have begun to articulate as a necessary individual and institutional strategy. The person who uses AI to do all their writing stops developing as a writer. The person who uses AI to research everything stops developing as a researcher. The question of what skills to maintain through deliberate human practice, rather than outsourcing to AI, is one of the most important personal and institutional decisions of the next decade. Almost nobody is making it consciously.
Keynes predicted fifteen-hour working weeks by 2030 and was wrong about the mechanism but possibly right about the destination. The economic problem — having enough — is closer to solved than it has ever been, for a growing share of humanity. The existential problem — what to do with yourself when the thing you organised your life around is no longer necessary — is more acute than it has ever been. We have no inherited wisdom adequate to the scale of what we are facing. We need to develop it, quickly, for real people on a real timeline.Neal Lloyd · Inside The Machine, Day 24
Inside The Machine, Day 24 · June 26 2026
Neal Lloyd writes about technology, human adaptation, and the uncomfortable questions nobody wants to answer at dinner. Inside The Machine is his ongoing daily series on AI.
- Day 01What Is This Thing?
- Day 02Survive the Machine
- Day 03The Great Debate
- Day 04Who Gets Hurt?
- Day 05Who’s In Charge?
- Day 06The Industries That Win
- Day 07The Human Edge
- Day 08The Creativity Question
- Day 09Does AI Feel Anything?
- Day 10The Data Problem
- Day 11The Trust Question
- Day 12The Accountability Gap
- Day 13The Rewired Brain
- Day 14Open vs Closed
- Day 15The New Cold War
- Day 16Why AI Lies With Confidence
- Day 17AI Is Eating the Power Grid
- Day 18The Age of AI Agents
- Day 19AI Safety Was Never Just Theory
- Day 20The Surveillance Question
- Day 21AI and the Future of Education
- Day 22AI and Your Health
- Day 23What Is AGI and Are We Close?
- Day 24What Is Work For?You are here
