Why UBI Is the Wrong Answer to the Right Problem

On artificial intelligence, labor displacement, and what we owe each other and the future

I. A crisis and an opportunity

Artificial intelligence will displace a significant number of jobs. This is not a distant forecast — it is already happening in paralegal work, financial analysis, mid-level management, customer service, and content production. The displacement is concentrated in precisely the segment of the workforce that least expected it: the educated, the white-collar, the people who were told that knowledge work was the safe harbor.

The dominant policy response being discussed is Universal Basic Income — a periodic unconditional cash transfer to every citizen, regardless of employment. The appeal is obvious. It is simple, it is dignified, it does not require anyone to prove need, and it addresses the most visible symptom of displacement: the loss of income.

This essay argues that UBI is the wrong answer. Not because the concern it responds to is wrong — the concern is entirely real — but because it addresses income while ignoring everything else that work provides. And it proposes, in place of UBI, something more ambitious: a deliberate redirection of displaced labor toward the accumulated backlog of things that decades of short-term thinking failed to build.

The disruption we are entering is real. But it is also, if handled deliberately, a rare window. Stable, prosperous societies do not reorganize. Disrupted ones do — but only if there is a coherent vision of what to reorganize toward. The argument here is that the vision exists, the need is genuine, and the moment is unusually well-suited to acting on it.

II. What work actually does

To understand why UBI falls short, you have to understand what work provides beyond a paycheck.

Work is structure. Without external deadlines, obligations, and routines, most people do not self-generate productive structure. This is not a moral failing — it is how human psychology works. We are social animals whose motivation is largely responsive rather than intrinsic. The research on self-determination theory, developed by Deci and Ryan over decades, establishes that even people with high intrinsic motivation benefit from external scaffolding. For people with lower autonomous motivation — which is the majority — that scaffolding is not a nice-to-have. It is load-bearing.

Work is identity. In modern Western societies, the first question strangers ask each other is: what do you do? This is not merely small talk. It is a probe for social location — a way of placing someone in a network of capability, contribution, and status. When that answer disappears, so does a significant part of how a person understands themselves and how others understand them. A 2018 study by Cassar and Meier in the Journal of Economic Perspectives established empirically what most people sense intuitively: work is a source of meaning that cannot be substituted by income alone. Paying someone not to work does not give them back what work provided.

Work is social fabric. For many adults, especially men, the workplace is the primary institution through which non-family relationships are formed and maintained. Remove the workplace and you remove the mechanism, not just the relationships. The resulting social isolation is not a minor inconvenience. The US Surgeon General’s 2023 advisory on loneliness classified social disconnection as a public health epidemic, associating it with mortality risk equivalent to smoking fifteen cigarettes per day.

Work is, finally, the primary mechanism by which people signal and receive confirmation that they are useful to others. This feedback loop — contributing something, receiving acknowledgment, contributing again — is fundamental to human wellbeing in a way that passive income receipt is not. A 2024 German study of UBI recipients (Malinka, Mitte, and Ziegler, in Springer’s Applied Research in Quality of Life) found that people with the lowest autonomous motivation gained the most from the income transfer — precisely because they were the most dependent on external structure. The implication is pointed: those most helped by UBI are those least equipped to replace what work provides in its absence.

UBI, in other words, is a solution designed by and for the minority of people who are highly self-directed, who already have strong social networks, who do not depend on their job for identity, and who would use unconditional time to pursue meaningful projects. That minority is real. It is also roughly ten to fifteen percent of the population. For the remaining eighty-five percent, unconditional income without structure does not produce flourishing. The evidence from joblessness makes this concrete.

III. What happens without work

We do not have to speculate about what mass joblessness produces. We have extensive empirical evidence from factory closures, industrial decline, early retirement, and long-term welfare dependency. The picture is not neutral.

Unearned income consistently predicts worse outcomes over time across populations that receive it. Cesarini, Lindqvist, and colleagues, in a landmark 2017 American Economic Review paper using data from Swedish lottery winners, found that windfall income significantly reduced labor supply — and produced mixed and often negative effects on wellbeing over time. The people who thrived were those with strong pre-existing social capital and self-direction. The people without those resources fared worse. Lottery winners, trust fund recipients, and long-term welfare recipients all tell variations of the same story: money solves the resource problem but leaves the meaning problem entirely untouched.

The gendered dimension of this is empirically robust and worth stating directly. Male joblessness correlates with antisocial outcomes — substance abuse, criminality, social withdrawal — in a way that female joblessness does not, at least not to the same degree or through the same mechanisms. A 2022 RAND study led by Shawn Bushway, published in Science Advances, found that the majority of unemployed American men aged thirty to thirty-eight had prior arrest records. A Harvard Business School study of mass layoffs in Colombia found a forty-seven percent increase in arrest probability in the year of job displacement, with effects persisting for years afterward, concentrated most heavily among young men. The mechanism is not simply financial desperation. It is the loss of routine, status, and the daily structure that keeps destructive behavior at bay. Money transfers do not address this mechanism.

The Harvard Colombia study added a particularly telling finding: when displaced workers received access to consumption credit — not cash grants, but credit — crime rates dropped significantly. The intervention that worked was not income replacement. It was the restoration of economic agency and forward orientation. People who could plan for the future behaved differently than people who received a passive transfer. This is not a subtle distinction. It suggests that the mechanism running from joblessness to destructive behavior is about meaning and agency, not only money — and that any solution which addresses only money will be incomplete.

IV. The strongest case for UBI — and why it still fails

The most serious argument for UBI is not that people will flourish in idleness. It is that UBI would function as a floor that enables risk-taking, retraining, and escape from work that is degrading. The person trapped in three simultaneous part-time jobs with no benefits, no security, and no time to develop themselves — UBI, on this account, would free them to retrain, start something, or simply stabilize. This is a real and meaningful argument.

The pilot data offers some support. Short-term UBI experiments in Kenya, Finland, and Germany all produced modest improvements in mental health and, in some cases, small increases in entrepreneurship. A 2021 systematic review of UBI pilots in high-income countries, published in Social Science and Medicine, found generally positive short-term mental health outcomes.

The problems are significant, however. First, all existing pilots are small in scale, temporary in duration, and limited to populations already experiencing poverty or precarity. None of them test the psychology of a society in which nobody is economically compelled to contribute. Temporary income relief and permanent unconditional income are fundamentally different phenomena — the first is a cushion, the second is a permanent reorientation of the relationship between individuals and economic necessity. The pilots tell us nothing useful about the second.

Second, meaningful UBI at scale has an inflation problem. When you give everyone unconditional income, you bid up the price of inelastic goods — housing, locally delivered services, food in urban areas. The real purchasing power gains erode as markets clear at higher prices. Asset owners capture the windfall. The people the policy was meant to help end up no better off in real terms while the cost has been enormous.

Third, and most fundamentally, the case for UBI as liberation assumes a model of human nature that fits a minority. The person who, given unconditional time and income, would pursue meaningful projects, develop skills, contribute to community, and live well — that person exists. But they were already navigating the world successfully. The person who drifts, loses structure, falls into dependency and isolation when external obligation is removed — that person is more common. A policy built on the psychology of the first group will disappoint when applied to the second.

V. The tree-planting failure

There is a Greek proverb that captures something essential about what a healthy society requires: a society grows great when old men plant trees whose shade they know they shall never sit in. The proverb describes an orientation — toward the future, toward people not yet born, toward investments whose return arrives long after the investor is gone. It is an orientation that modern institutions systematically violate.

The economic term for this failure is hyperbolic discounting: humans systematically overweight the present and undervalue the future, at a rate that exceeds what rational actors would. This tendency is amplified by the institutional structures we operate within. Politicians face re-election every four years. Corporations report earnings every quarter. Neither has a structural incentive to plant trees whose fruit arrives in twenty years. The result is a systematic accumulation of deferred costs — costs that are clearly visible in advance, that experts consistently flag, and that are consistently kicked down the road anyway because the pain of addressing them falls on the present while the benefit accrues to the future.

We have been living inside this failure for decades. The bill is now visible.

Physical health: metabolic disease — obesity, type 2 diabetes, cardiovascular disease — has risen dramatically across developed nations since the 1980s. The causes are substantially structural: sedentary work, urban environments designed for cars rather than bodies, food systems optimized for profit rather than nutrition. These are not individual failures of willpower. They are the predictable outputs of environments engineered for short-term efficiency without regard for long-term health. The healthcare costs this generates are enormous and still accelerating.

Mental health: Jonathan Haidt and Jean Twenge’s research documents a sharp deterioration in adolescent mental health beginning in the early 2010s, with anxiety, depression, and loneliness rising steeply across age groups, accelerating before COVID and continuing after. The causes are structural, not individual. A generation has been produced with measurably higher baseline distress than their parents, partly through the environments and incentives we built around them.

Infrastructure debt: roads, bridges, water systems, and electrical grids are being run to failure across much of the developed world. Flint, Michigan’s water crisis is the visible version of a problem that exists at lower visibility in thousands of similar systems. Maintenance generates no political capital, so it is deferred. Deferred maintenance always costs more when it eventually fails — but the failure arrives on someone else’s watch, so the incentive to defer remains.

Intergenerational equity: each successive generation in most developed nations faces higher housing costs, more expensive education, lower upward mobility, and a larger environmental and fiscal debt than the one before. The generation currently entering the workforce will be the first in modern history to be materially worse off than their parents across most of these dimensions. This is not accidental. It is the direct consequence of a series of decisions that maximized present consumption at the expense of future conditions.

Social capital: Robert Putnam documented in Bowling Alone the 40-year decline in civic participation, community institutions, and interpersonal trust in the United States. The research has been replicated and extended across most developed nations. The trend has continued and, by some measures, accelerated. Social isolation is now classified by the US Surgeon General as a public health epidemic with mortality consequences comparable to smoking. We built environments and economic structures that progressively dissolved the social fabric — not because anyone intended this, but because social capital does not appear on quarterly earnings reports or four-year political ledgers.

Elder care: aging populations were demographically predictable for decades. The infrastructure to care for them was not built. In most developed nations, elder care is addressed either by burning out unpaid family members — disproportionately women — or by expensive private facilities that most families cannot afford, with very little in between. The crisis was visible. The tree was not planted.

None of these are surprises. Every one of them was flagged in advance by researchers, public health officials, engineers, and demographers. Every one was deferred by institutions optimized for the short term. The common mechanism is identical across all of them: the cost falls on future people; the benefit of deferral accrues to present people and institutions; the future people had no vote. This is the tree-planting failure operating at civilizational scale.

VI. The long-term cost of short-term efficiency

This is where the two problems fuse. The displacement described in the preceding sections — the loss of structure, identity, and social fabric — and the infrastructure backlog described above are not separate crises requiring separate policy responses. They are, if you look at them together, a lock and a key. The labor that AI will displace is precisely the kind of structured, purposeful, socially embedded work that human psychology requires. The backlog of things that short-term thinking failed to build is precisely the kind of large-scale, legible, durable work that can provide it. The argument is not merely that we should redirect displaced workers toward infrastructure as a pragmatic stopgap. It is that building the infrastructure is the psychological intervention — the restoration of structure, contribution, and forward orientation that income transfers cannot provide.

Here is the standard economic objection to large-scale public works programs: they are inefficient. We are paying people to build things before the market has priced the need. The resources could be allocated more productively. The argument sounds rigorous, and in a narrow time horizon it is. But it is the same short-term logic that produced the accumulated failures described above, and it makes the same mistake: it zeros out the cost of the counterfactual.

Idle, purposeless people are not a neutral economic state. They generate costs that do not appear on the short-term efficiency ledger: healthcare — addiction, depression, metabolic disease from sedentary and structureless lives — criminality, political radicalization, and the generational transmission of dysfunction to children who grow up watching their parents without purpose or direction. These costs are real, they are large, and they compound. A society that tolerates mass purposelessness in the name of short-term fiscal discipline is not saving money. It is moving the bill to a future ledger where it will arrive larger, less legible, and harder to address.

In Keynesian economics, there is a well-established argument that in demand-slack economies — precisely the kind AI disruption would create — infrastructure investment generates larger multiplier effects than transfer payments. Transfer payments produce consumption. Infrastructure investment produces consumption plus durable capital: the roads, grids, housing, and care systems that generate value for decades after the initial investment. The New Deal’s Civilian Conservation Corps built roads, bridges, and parks still generating economic and social value ninety years later. The argument for infrastructure over transfers is not ideological. It is about what you have left when the spending is done.

The power infrastructure case makes this argument in its purest form, because it dissolves even the “premature” objection. AI data center electricity demand in the United States alone is projected by Deloitte to grow more than thirtyfold by 2035, from four gigawatts to one hundred and twenty-three gigawatts. The National Center for Energy Analytics estimates that seventy-five to one hundred gigawatts of new generating capacity will be needed by the early 2030s — and over seventy-five gigawatts are already permitted or under construction. Any grid capacity built now is not excess. It is early. The need is not speculative. It is in construction pipelines. Jensen Huang described this at Davos in January 2026 as “the largest infrastructure buildout in human history.” Nvidia, Microsoft, Amazon, Google, and Meta have collectively committed hundreds of billions annually to the physical infrastructure of AI. The labor to build it is already in severe shortage.

The choice, framed correctly, is not between productive investment and wasteful make-work. It is between building things slightly ahead of when the market would demand them — while solving a social crisis right now — versus the alternative: paying the social costs of mass purposelessness through healthcare systems, prisons, and political instability over the next thirty years, while also eventually having to build the infrastructure anyway, under worse conditions, at higher cost. Excess production in infrastructure is a better bet than optimal short-term efficiency combined with long-term social deterioration.

VII. Quality of life is not negotiable

There is an assumption buried in most discussions of labor market transition that deserves to be named and rejected: that people are infinitely substitutable and psychologically resilient to sudden status loss. They are not.

A person who spent fifteen years building a career as a financial analyst, a paralegal, or a mid-level manager — who arranged their mortgage, their children’s schooling, their sense of self, and their social standing around that career — did not consent to having it dismantled by a technological shift they did not cause and could not have anticipated. Telling them that UBI will cover their basic needs is not a transition. It is a status collapse with no narrative attached. And the combination of purposelessness and perceived degradation — not just poverty, but the experience of having had something and lost it — is historically where resentment concentrates. It is the psychology not of acceptance, but of political volatility.

Any serious transition program must meet people where they are, not at a subsistence floor. This has concrete implications for how labor redirection is designed.

First, displaced knowledge workers should not be defaulted to manual labor. Their skills in logistics, planning, data analysis, compliance, and coordination are genuinely needed in large infrastructure programs — in project management, regulatory navigation, procurement, quality assurance, and operations. The transition should identify where existing skills transfer, not assume they do not.

Second, wages need to allow reasonable continuity of life. Not luxury — but mortgages, children’s education, healthcare. Retraining stipends and transitional employment that cover only subsistence are not a dignified transition. They are humiliation with better optics than UBI.

Third, framing matters enormously. “You are building the power grid for the AI economy” is a fundamentally different psychological offer than “here is a public works position because we do not know what else to do with you.” The difference is not cosmetic. It is the difference between contributing to something and being warehoused by something. The closest historical template is not the welfare state. It is the Second World War’s industrial mobilization — the rapid conversion of a peacetime workforce into defense production, wages preserved, social cohesion maintained, a national narrative of contribution that made the disruption legible and even ennobling. People can absorb enormous disruption when they understand what it is for.

VIII. What to build — and why it is not make-work

The infrastructure case is stronger than most of its advocates have stated, because it connects to almost every one of the accumulated deferred problems described earlier. These are not arbitrary public works projects. They are the backlog of things we failed to build during the decades of short-term thinking — and most of them are also exactly what the AI economy will need.

Power and AI infrastructure is the most immediate. The construction industry is already short approximately four hundred and thirty-nine thousand skilled workers, according to the Associated Builders and Contractors. The Bureau of Labor Statistics projects electrician demand growing nine percent through 2034, faster than nearly all other occupations. BlackRock launched a hundred-million-dollar Future Builders Initiative in March 2026 specifically to address the shortage of electricians, pipefitters, steelworkers, and network technicians needed to build out AI data centers, grid modernization, and energy systems. These are well-paid, learnable, durable jobs that AI will not automate. They are also, right now, unfilled.

Housing addresses the intergenerational equity failure directly. Decades of undersupply in most major cities have produced affordability crises that now reach across income levels. The waiting lists for social housing in many European cities are measured in decades. In North American cities, the ratio of median home price to median income has reached levels that effectively bar younger generations from ownership. This is labor-intensive work with immediate social value and a visible output: people can see what was built and who lives in it.

Health infrastructure — walkable urban environments, parks, community recreation facilities, cycling networks — addresses the metabolic and mental health crises that sedentary modern life has produced. These are not luxuries. The research on built environment effects on physical activity, mental health, and social connection is extensive and robust. A park built in a food desert is preventive medicine at scale, cheaper across a population than the diabetes, cardiovascular disease, and depression it forestalls. Green spaces reduce cortisol. Walkable neighborhoods reduce isolation. This is not speculative.

Elder care addresses a crisis that was demographically predictable for thirty years and consistently deferred. It is deeply labor-intensive — a sector where human presence is not merely efficient but essential, and where AI substitution is both technically and socially limited. It also carries a particular meaning: how a society treats its elderly is legible to younger generations as a signal about what they can expect. Building the infrastructure of elder care is one of the clearest expressions of the tree-planting principle available — it is investment in people who cannot contribute economically, on behalf of a society that understands its obligations extend beyond the immediately productive.

Environmental repair — reforestation, soil health, clean water systems — offers the slowest return on this list, and is for that reason the purest expression of the argument. Trees planted now will provide shade to people not yet born. Aquifers recharged now will supply water to communities two generations away. This is not inefficiency. This is what the Greek proverb describes. The disruption that is forcing this reorganization could, if we choose, be the moment we finally start planting seriously.

IX. A crisis worth not wasting

There are two futures available from this disruption.

In the first, we respond with UBI — or with nothing deliberate at all. People receive income transfers and lose the structure, identity, and social fabric that work provided. The behaviors that emerge from mass purposelessness compound over a generation: declining health, rising criminality, eroding social trust, political radicalization, and the intergenerational transmission of disorientation to children who grow up without the model of purposeful contribution. This is not conjecture. It is an extrapolation from what we already know happens when communities lose their economic purpose — from deindustrialised towns in the American Midwest and North of England, from regions where employment collapsed and was never replaced. The outcomes are consistent and grim. UBI would provide slightly more comfortable circumstances for the same trajectory.

In the second, we treat the disruption as the forcing function that finally generates the political will to address what stable prosperity never could. We redirect the labor that AI displaces toward the backlog of things we should have built decades ago: the power grid that the AI economy will require, the housing that the next generation cannot afford, the health infrastructure that sedentary modern life has made necessary, the elder care facilities that aging populations need, the green spaces and water systems and community institutions that social cohesion depends on. We do this at wages that preserve dignity, through programs that find where existing skills transfer, with a narrative that is honest about what is happening and why.

There is a version of this transition in which a generation looks back and says: we built the infrastructure of the next economy. We fixed the water systems. We planted the forests. We created the housing. We cared for the elderly. The disruption was real, but we used it.

Once the transition stabilizes — once the infrastructure exists, the new labor distribution is established, and the economy has absorbed the shock — market mechanisms can reassert themselves. The intervention is transitional, not permanent. But the infrastructure built during it is permanent. The social capacity maintained during it is permanent. The generation that came through with purpose and dignity rather than subsidised purposelessness is permanent in the sense that matters most: its children will be different.

The question is not whether we can afford to redirect labor toward the things we have been too shortsighted to build. The question is whether we can afford the alternative — and whether we have the imagination to see a crisis as an opportunity before it becomes only a catastrophe.

That is the difference between planting trees and handing out fruit.