Universal Basic Income and AI: Will the Robots Pay Us to Live? (An India Reality-Check)

Every few months, some tech founder floats the same soothing promise: don’t worry about the machines taking your work, because a universal basic income — AI-funded, no strings attached — will catch everyone who falls. The pitch is seductive precisely because it sounds like justice. If artificial intelligence is going to absorb the labour that used to feed a family, then let the wealth it generates flow back to the people it displaced. A monthly floor under every citizen, machine-financed, unconditional. Who could argue against being paid to live? But the closer you look at the idea — and especially at what it would mean in India — the more you notice that the promise quietly skips the hardest question of all: who owns the machine doing the paying?

The pitch: if AI eats the jobs, pay everyone a floor

The argument for a basic income in the age of AI is genuinely strong, and it deserves to be heard on its best terms rather than caricatured. It goes like this. Automation has always destroyed particular jobs while creating others, but this wave feels different — it reaches into cognitive work, into the white-collar and services roles that were supposed to be safe. If a large share of tasks can be done more cheaply by software, wages for ordinary people come under sustained pressure, and the returns pool with whoever owns the models and the data centres. A basic income breaks that spiral: give every adult a regular, unconditional cash payment, enough to cover the essentials, and you decouple survival from a labour market that may no longer need as many hands.

There is dignity in the design. Cash respects the recipient’s judgement — a mother knows better than a bureaucrat whether this month’s rupees should go to school fees or medicine. It cuts the paternalism and leakage of sprawling welfare schemes. And it is, at least in theory, a clean answer to a frightening future: if you are anxious about whether the coming years will leave a place for you, as many readers of will AI take my job are, a guaranteed floor sounds like the one policy that makes the anxiety bearable. This is not a fringe fantasy. India’s own Economic Survey of 2016–17 devoted a serious chapter to universal basic income as a possible replacement for a tangle of subsidies, and several Indian states already run direct cash-transfer schemes for women and farmers. The idea is on the table. The question is what it actually costs, and what it actually changes.

The India math: why a nationwide liveable UBI is a near-impossible sum today

Start with arithmetic, because this is where the dream meets the ground. A UBI in India that is both universal and liveable runs into a wall of scale. India has well over a billion people. To be meaningful, a payment has to be large enough to actually matter to a poor household — not a token, but a genuine floor. Multiply any such figure, even a modest monthly one, by a billion-plus recipients, twelve months a year, and you arrive at a number that swallows a very large share of national output.

I want to be careful here and reason in orders of magnitude rather than pretend to a precise figure. A payment set low enough to be affordable stops being liveable; a payment set high enough to be liveable stops being affordable. That is the trap. Economists who have modelled versions of this for India tend to land in the same uncomfortable place: a genuinely universal, genuinely liveable transfer would cost a share of GDP so large that it could only be funded by either dismantling most existing subsidies and welfare, or by taxation on a scale the country does not currently raise. India’s tax-to-GDP ratio is modest by rich-world standards, and much economic activity sits in the informal sector beyond easy reach of the taxman. You cannot redistribute revenue you are not collecting.

A payment set low enough to be affordable stops being liveable; a payment set high enough to be liveable stops being affordable. That is the trap.

This is why the honest versions of the Indian debate rarely propose a full, rich-country UBI. They propose something smaller and quasi-universal — a “basic income” that is really a better-targeted, less leaky cash transfer, often aimed at women or the rural poor, and often explicitly traded off against existing subsidies. That may well be good policy. But notice what has happened: the machine-financed, live-anywhere, work-optional floor of the AI pitch has shrunk into a means-tested top-up. The word “universal” is doing enormous work it cannot actually afford. And the promise that AI’s bounty will fund it assumes both that the bounty is enormous and that the state can capture it — two assumptions that deserve far more scrutiny than they usually get.

Universal Basic Services and “participation income” as alternatives

If a full cash UBI is fiscally out of reach, the debate does not end — it forks. Two alternatives are worth taking seriously, because they attack the same problem from different angles.

The first is Universal Basic Services. Instead of handing people money to buy essentials in a market, the state provides the essentials directly and freely at the point of use: healthcare, schooling, public transport, water, perhaps digital connectivity and a floor of nutrition. The logic is that a rupee of public provision often stretches further than a rupee of cash, because the state can buy at scale and because some goods — a functioning clinic, a school — are cheaper to guarantee collectively than to purchase individually. For a country still building out basic infrastructure, strengthening what already exists — the public health system, the midday meal, the ration network — may simply deliver more security per rupee than a cash floor ever could. The catch is that services can be captured, degraded, or delivered badly, and cash gives people a freedom that a queue at a broken clinic does not.

The second alternative is a participation income, sometimes framed as a job guarantee. Here the payment is conditional not on poverty but on contribution — caring for a relative, studying, doing community or ecological work, or taking guaranteed public employment. India already runs the world’s largest experiment in this direction through its rural employment guarantee, which offers a legal right to a fixed number of days of paid work. A participation income tries to keep the dignity of contribution while providing the floor, and it sidesteps the political objection that unconditional cash pays people to do nothing. Its weakness is administrative: conditions mean bureaucracy, and bureaucracy means gatekeepers, exclusion errors, and exactly the leakage a clean UBI was meant to avoid.

These are real, substantive choices, and reasonable people land in different places on the basic income pros and cons. But all three options — cash, services, participation — share a blind spot, and it is worth naming plainly.

The catch: a payout doesn’t undo who owns the machine

Here is the question that the UBI pitch is engineered to make you forget. Suppose it all works. Suppose AI does generate enormous wealth, and suppose the state manages to tax enough of it to send every citizen a monthly payment. Who still owns the models, the data, the compute, the platforms? The same handful of firms and individuals who owned them before. The cheque changes what you receive. It does not change who decides.

This is a pattern that runs across ten thousand years — the same move, a new machine, every time. A new technology concentrates power; the people it displaces protest; and the owners buy peace by handing back a slice of the gains while keeping the engine itself. A basic income, in this reading, can become the most elegant version of that bargain ever devised: a permanent, monthly settlement that funds your consumption precisely so that you never contest the ownership. You become a customer of the machine economy, paid to stay one, with no seat at the table where its rules are written. That is not liberation. It is a very comfortable form of dependence.

The cheque changes what you receive. It does not change who decides.

None of this means a basic income is bad. A hungry family needs cash now, not a lecture about ownership, and a floor that prevents destitution is worth having on its own terms. The point is narrower and sharper: a payout is a distribution of output, not a distribution of power. If the entire political imagination of the AI age reduces to negotiating the size of everyone’s allowance, then the people who own the machines have already won the more important argument — the one about who gets to own them. Real answers look different. They look like platform cooperatives where workers hold the algorithm, like public stakes in critical models and data, like the patient organising behind how India’s farmers won their concessions. Those change who decides. A cheque, by itself, does not.

Technical fixes vs political power — the crypto cautionary tale

There is a deeper trap here, and the last decade handed us the perfect demonstration of it. When cryptocurrency arrived, it promised to decentralise money itself. No banks, no central authority, power distributed across a network — the technology was, genuinely, decentralised by design. And yet within a few years the ecosystem had reconcentrated with startling speed: a few giant exchanges, a few whales holding the bulk of the coins, mining pooled into a handful of operations, and ordinary users once again dependent on intermediaries they had to trust. The ledger was decentralised. The power was not.

That is the distinction that matters most for the UBI debate: technical decentralisation is not the same as political decentralisation. You can build a system that is beautifully distributed at the level of code and still have it deliver control into very few hands, because ownership, capital, and coordination reassert themselves through every gap you leave open. Crypto decentralised the plumbing and centralised the wealth. A basic income risks the mirror image: it distributes the cash and leaves the ownership of the machines exactly where it was — and calls that fairness.

So the right way to read “will AI lead to UBI” is not as a yes-or-no forecast but as a warning about framing. It very well might lead to some form of basic income — the owners of the technology may find it the cheapest way to keep the peace. The question a thoughtful reader should ask is not only how big is the cheque but who signs it, who can stop signing it, and who owns the thing that makes the money in the first place. A UBI that arrives as a gift from the machine’s owners is a gift they can revoke, condition, shrink, or weaponise. A floor that citizens hold through genuine political power — through ownership, through law, through organisation — is something else entirely.

Pay everyone to live, by all means. Prevent destitution, build the services, guarantee the floor. But do not mistake the allowance for the answer. The robots, if they pay us at all, will pay us with money that belongs to whoever owns them — and until that ownership is on the table, the same old move is simply wearing a newer, kinder mask. That is the reality-check India, and the rest of us, cannot afford to skip.