If artificial intelligence is squeezing out entry-level work and the pension system is heading for collapse, what will Generation Z live on? A universal basic income (UBI) — a fixed sum paid to every citizen each month, no questions asked and no conditions attached — has been debated for decades. Now that several trials have reported results, it deserves to be taken more seriously. This instalment looks at what the data show: does a basic income work, and if it does, how can it be paid for?
1. The core idea — what is a universal basic income?
A universal basic income (UBI) — whose principles and promises an earlier analysis examined in detail — rests on a simple idea: every citizen receives a regular, fixed cash payment simply by virtue of birth, residence or citizenship. There are no conditions: no requirement to work, no requirement to look for work, no means test. The aim is not simply to abolish poverty — although that is one of the effects — but to build a safety net that lets people change their lives, change careers or retrain without going hungry in the meantime.
The idea is not new — something close to it appears in Thomas More's Utopia of 1516 — but advances in artificial intelligence and automation have given the debate a fresh urgency. Its history runs from Thomas Paine to Martin Luther King; technological displacement has brought it back into public debate. In his 1967 book Where Do We Go from Here: Chaos or Community?, King put it plainly: "the solution to poverty is to abolish it directly by a now widely discussed measure: the guaranteed income." Yet the idea was never implemented at scale until the 2010s, when technological change put it back on the agenda.
If artificial intelligence squeezes out routine work, and fewer and fewer people are in stable, conventional employment, then the traditional model of earning a living (work → wage) is called into question. Seen in that light, a UBI is not "free money" but a structural response to a structural problem created by technological progress.
A universal basic income (UBI). Think of it as a citizen's dividend. Just as a company's shareholders receive an annual dividend out of its profits, a UBI means that every citizen receives a fixed sum out of the country's economic output — not because they have worked for it, but because they are a member of society. The point is that there are no conditions. There is no requirement to be unemployed, no requirement to be poor, nothing to prove to anyone. The aim is that everyone should have a floor to build on — whether that means retraining, starting a business, or simply the security of getting by.
2. The big experiments — what do the data show?
In recent years, basic-income trials have been run in a number of countries and settings. These are not theoretical models but studies built on measurable data, firm enough to draw conclusions from.
OpenResearch / Sam Altman's experiment (2020–2024)
The largest and longest-running American basic-income trial was run by OpenResearch — the non-profit research arm of OpenAI, the company behind ChatGPT — and funded by Sam Altman. It tracked 3,000 participants over three years: one group received $1,000 a month, the control group $50. The payments began at the end of 2020, during the COVID-19 pandemic and the extraordinary labour-market shutdowns that came with it — which the researchers regard as an important limiting factor. The results were published in July 2024.
The central finding: recipients did not leave the labour market. The employment rate fell by 2.0 percentage points relative to the control group — statistically significant, but a small decline. Recipients did, however, cut their hours (by 1.3 a week on average), and the largest effect was greater financial breathing space: less stress, and more time for family, for study and for starting a business.
Finland's experiment (2017–2018)
The Finnish trial followed 2,000 unemployed participants, who received €560 a month for two years. The results were published in 2020. The employment effect was slight: over the course of the year, recipients worked on average six days more than the control group — but the difference was not statistically significant.
The real effect lay in mental well-being. Recipients were more satisfied with their lives than the control group, and were less likely to report depression, sadness or loneliness; they also rated their own cognitive functioning — concentration, learning, decision-making — more highly. Participants said that sense of security let them plan further ahead instead of focusing on day-to-day survival.
| Experiment | Duration | Participants | Payment | Employment effect | Other effects |
|---|---|---|---|---|---|
| OpenResearch (USA) | 3 years | 3,000 | $1,000 a month | No withdrawal from the labour market (-2.0 pp employment, -1.3 hours a week) | Financial breathing space, less stress |
| Finland | 2 years | 2,000 | €560 a month | Slight (+6 working days a year, not statistically significant) | Less depression and stress, better self-reported cognitive function |
| Stockton SEED (USA) | 2 years | 125 | $500 a month | Year 1: +12 pp in full-time work; final result not statistically significant (pandemic) | Less stress, more spending on health |
| GiveDirectly Kenya | 12+ years (study period) | 26,500+ households | $1,000 lump sum | Did not stop working | 2.5× economic multiplier, neighbours gained too |
3. The economic effects — beyond "free money"
Critics of a basic income often argue that handing out money causes inflation and destroys the incentive to work. The experimental data partly refute that argument and partly qualify it.
GiveDirectly's Kenyan trial — which gave a one-off transfer of $1,000 to more than 26,500 households — had a surprising economic effect: every dollar paid out generated $2.50 of local economic activity. That "multiplier effect" arose because participants spent the money locally — they bought grain at the local shop, the shopkeeper paid the local mill, the mill took on more workers, and so on. Neighbours who received no transfer gained as well: within a year, their income and consumption had come close to those of the recipients.
The inflationary effect in the trials was slight. GiveDirectly's research found that local prices rose only marginally, because local supply easily kept pace with the extra demand — shopkeepers ordered more stock and producers increased output. That suggests a basic income need not be inflationary in itself, provided local supply is elastic enough to keep pace with the demand it creates.
The labour-market effects, though, are not unambiguously positive. In the Finnish trial, recipients did not leave the labour market — but nor did they enter it in greater numbers. A basic income, in other words, cannot solve employment problems on its own. Creating jobs still requires an environment conducive to enterprise, training programmes and market demand. In that context, a UBI is better understood as a safety net that makes it possible to take risks: if people know they have a basic income to fall back on, they are more willing to start a business, take a course, or retrain for a new trade. In the Kenyan trial, considerably more small businesses were started with the money — but these were typically local, small-scale ventures that do not necessarily create sustainable, long-term jobs.
Longer-term economic modelling raises a further question: what happens if a basic income is not an experiment but a permanent system? The decisive question is how it would be financed. Work by the Nobel laureates Abhijit Banerjee and Esther Duflo and their research group suggests that financing a UBI in an advanced economy would require a substantial tax rise: 10–15% of GDP would have to be redirected. That is exceptionally difficult politically, particularly in countries where the tax burden is already high. Nor can the Kenyan model be transferred directly: in an advanced economy, the local multiplier effect is weaker, because a larger share of consumption goes on imported goods and local productive capacity is less elastic.
4. Alternative models — a UBI is not the only option
A basic income is not the only answer to technological displacement and the crisis of the pension system. Several alternative models approach the problem in different ways:
A negative income tax (NIT): Milton Friedman's idea. Those whose income falls below a given threshold do not pay income tax — they receive a payment through it. It is not for everyone, only for those on the lowest incomes — which makes it cheaper than a UBI, but does less to simplify the welfare system.
A conditional basic income: recipients have to meet certain requirements — studying, volunteering or looking for work, for instance. This model preserves "work incentives", but loses one of the UBI's most important advantages: freedom from bureaucracy.
A robot tax: Bill Gates's proposal — companies using artificial intelligence and robots would pay a special levy, and that levy would fund the basic income. This model ties the gains from technological progress directly to supporting those displaced by it.
Universal basic services: under some proposals, the answer is not to hand out money but to provide the basic necessities — housing, healthcare, education, transport — free of charge or very cheaply. This model is less flexible than a cash transfer, but it guarantees that the support meets those needs.
| Model | Who receives it? | Conditions? | Advantage | Drawback |
|---|---|---|---|---|
| UBI (unconditional) | Every citizen | None | Simple, no bureaucracy, no stigma | Expensive; the rich get it too |
| Negative income tax | Only those on low incomes | An income threshold | Cheaper, targeted | Bureaucracy, income testing |
| Conditional | Participants | Study/volunteering/job-seeking | Preserves "work incentives" | Bureaucracy, monitoring the conditions |
| Robot tax | Every citizen | None | Funded by AI | Corporate resistance, tax avoidance |
| Universal basic services | Every citizen | None | Guaranteed to meet real needs | Inflexible, bureaucratic |
5. The Hungarian context — is any of it realistic?
In Hungary, a basic income is not on the agenda. The state of the public finances — high government debt, the suspension of European Union funds and rising pension payments — rules out introducing a transfer programme on that scale. A UBI of 150,000 forints a month, confined to the adult population (around 7.4 million people), would cost 13.32 trillion forints a year — 40% of the entire annual budget. If the income were genuinely universal, applied to the full population of 9.6 million, the annual bill would come to 17.28 trillion forints. An earlier analysis — the Hungarian test — followed three real Hungarians, Gábor, Mária and Dávid, to show who would gain from a basic income and who would be worse off.
The structural problem, though — displacement in the labour market, an ageing society and an unsustainable pension system — is not going away. If current trends continue, the Hungarian government will sooner or later have to face the fact that the traditional, work-based model of income security no longer works. In that situation, a basic income — or one of the alternative models — may be less a "luxury" than a necessity.
The question, then, is not whether Hungary will have a basic income, but when and in what form. Generation Z — caught between AI displacement and the collapse of the pension system — may be the first to face that question in practice.
In summary: the basic-income experiments show that people do not stop working when they are given a basic income — but the kind of work people do, and what motivates them, changes. In the Finnish trial, depression fell; in the American trial, financial pressure eased; in the Kenyan trial, the local economy took off. A basic income is no magic wand — it does not solve the demographic crisis, it does not substitute for structural reform, and it cannot be funded without limit. But in a world where technology makes ever more people "redundant" in the traditional labour market, a basic income may be one of the instruments that prevents society from splitting in two: those in work and those without it. The next instalment turns to what can be done on the other side of the problem: how Big Tech might be regulated, and whether accountability can be enforced.
Recipients did not leave the labour market. The largest effect was that they could devote more time to their families, their studies and their own businesses.
— Finding of the OpenResearch Unconditional Cash Study (2024)
Every dollar transferred generated $2.50 of local economic activity. The money did not disappear — it circulated, and everyone gained from it.
— Finding of the GiveDirectly transfers in Kenya (2022)