From left: legislator Ko Ju-chun (葛如鈞), UBI Taiwan chair Jiakuan Su (蘇嘉冠), and standing director Henry Lee / Li Pinyi (李品逸) recording EP350. Image reproduced from the original BIEN report.
A podcast episode is easy to overread. Politicians host guests. Advocates gain an audience. Everyone posts a photograph, and public policy moves on.
EP350 of legislator Ko Ju-chun’s podcast, Dr. Bao Talks with Friends, deserves closer attention because it is part of a bigger trend. Ko has now devoted three episodes to basic income. UBI Taiwan has met with him repeatedly inside the Legislative Yuan. He has discussed the issue with Basic Income Earth Network chair Sarath Davala. His team is now preparing to present its own Universal Basic AI proposal alongside UBI Taiwan at the 2026 BIEN Congress in Toronto.
The important development is therefore not a politician saying something friendly about basic income. It is the creation of a working relationship with a calendar, policy prototypes, and a route into both Taiwan’s legislature and the international basic-income movement.
That matters because Taiwan’s basic-income debate is entering through a door that advocates often overlook. It is not entering mainly through a conventional argument over welfare budgets. It is entering through artificial intelligence, semiconductors, child accounts, financial technology, and the growing gap between Taiwan’s national success and the economic insecurity many people experience at home.
A relationship built across several conversations
Ko is an unusual political interlocutor for a basic-income organization. Before entering the Legislative Yuan as a Kuomintang (KMT) legislator, he worked across technology entrepreneurship, digital media, design, and academia. He approaches UBI from the world of technological transition rather than from Taiwan’s established social-welfare institutions.
UBI Taiwan’s first episode on his program, EP277, introduced unconditional basic income to his audience through Jiakuan Su and Henry Lee. In October 2025, Ko recorded EP313 with BIEN chair Sarath Davala and me, asking whether an idea associated with Europe and North America could develop a distinctly Asian path.
Ko has noted that Taiwan is a digitally connected society of manageable scale, people have now experienced repeated government vouchers and universal cash payments, and the fear surrounding a national cash experiment is falling.
He also made an observation that has become more important with time. Taiwan’s AI and semiconductor economy can produce extraordinary aggregate numbers while young people feel little improvement in wages, housing access, or career security. Ko described how rival parties fought over the NT$10,000 payment and the government ultimately delivered it. Once every major camp has participated in universal cash, the policy is harder to dismiss as the property of one ideology.
An important development is the proposed National Account for children that Ko and others pushed. Taiwan is no longer debating UBI as a distant endpoint, but as something appearing in bits and pieces across multiple policies.
This background changes how the episode should be read. Ko is not endorsing a finished UBI bill. He is doing something more useful at this stage: treating basic income as a legitimate technology-policy question and helping move it between media, parliament, civil society, and international networks.
Taiwan is sitting on the next generation’s oil field
The most memorable line in EP350 came from Henry Lee. Oil shaped the economic and geopolitical order of the last century, he argued. The next century will be built around silicon, artificial intelligence, and compute. Taiwan’s position in advanced semiconductors means that future Taiwanese will be “born on the oil field of the next generation.”
The metaphor works because it contains both pride and danger.
A country can sit above an oil field while most residents own none of it. Resource wealth can finance public institutions and household security, or it can concentrate power around companies, asset owners, and political insiders. The same choice now faces Taiwan. The island supplies a foundation of the global AI economy, yet much of the gain reaches households through channels they do not equally possess: technology salaries, stock ownership, business equity, and property.
This is why the old automation argument was too narrow. Basic-income advocates often warned that robots would eliminate jobs and eventually force the state to replace wages. That prediction created an unnecessary all-or-nothing test. If headline unemployment stayed low, critics could declare the problem imaginary.
AI changes economic power before it creates a dramatic unemployment number. It can erase junior tasks without eliminating an entire occupation. It can allow one experienced worker to do the work once assigned to several newcomers. It can narrow the first rungs of a career ladder, intensify monitoring, and raise output expectations. This is harder to read in the data which makes it even more insidious.
Ko has said that young people may understand AI tools, yet experienced workers can learn those tools faster than a new graduate can acquire years of professional judgment.
That is a different threat from mass joblessness. A society may retain high employment while making entry, mobility, and bargaining power steadily worse. The worker still has a job, but fewer credible alternatives. The graduate still has a degree, but fewer routes to build experience. Productivity rises while the value of being replaceable falls.
Ko has already carried part of this concern into legislation. He has pointed to the labor provisions in Taiwan’s AI Basic Act, which direct government to close skill gaps, support labor participation, protect economic security, and pursue decent work. Those commitments are meaningful, but retraining and job placement still assume that the central task is keeping people employable. Basic income addresses the part a labor statute cannot: whether a person has bargaining power during a transition, enough time to acquire a new skill, and a floor when the available job is damaging or poorly paid.
Taiwan’s AI debate therefore needs to ask two questions at once. Who can use the technology? Who owns the economic leverage it creates?
Universal Basic AI answers only the first question
Ko’s most original contribution to this debate is Universal Basic AI. His team is developing the idea further as UBAI: AI Coin, a public entitlement that would allow each person to obtain a baseline amount of high-quality AI service or compute. The proposal draws a useful distinction between financial inequality and intelligence inequality. If premium models, specialized agents, and reliable compute become prerequisites for education, work, healthcare, and participation, unequal access will reproduce class advantage.
A worker without AI access may soon resemble an office employee who cannot use basic software. A student with a strong private tutor-model may compound an advantage over a student limited to a weak free tier. A small organization may be locked out of capabilities that a large company treats as ordinary infrastructure.
Taiwan should build a capability floor: affordable access to strong models, local-language systems, compute, training, privacy protection, and a meaningful ability to move one’s data and work between providers.
Yet access alone can become a sophisticated way of adapting people to an unequal economy. Giving everyone a better tool does not determine who captures the value produced with it.
Lee posed the question well during the episode: will human–AI cooperation become universal, or will it become elite? He sketched a deliberately uncomfortable future in which a small group controls the most powerful productive systems while most people use AI to discuss their feelings, analyze horoscopes, or watch synthetic short dramas. The point was not contempt for ordinary uses. It was that access to an interface can coexist with exclusion from ownership, revenue, and decision-making.
The distinction is simple. Universal Basic AI distributes productive capability. Basic income distributes independent purchasing power. A public wealth dividend distributes ownership income. Each instrument solves a different problem.
Compute credits cannot pay rent. A model subscription does not cover childcare during retraining. An AI tutor cannot give a worker the financial room to leave an abusive job. In fact, equal access to AI could intensify competition if every worker is expected to produce more while firms retain the savings.
Taiwan should pursue a dual dividend: a capability dividend that guarantees access to the tools of the AI economy, and an economic dividend that gives every person a direct share of the prosperity built around those tools. One without the other leaves the social contract incomplete.
The family experiment made the argument concrete
EP350 also returned to UBI Taiwan’s two-year basic-income experiment with a single mother and her son. The participant received NT$10,000 per month. The amount did not transform her into a different person or remove every difficulty from her life. It changed the margin within which she could make decisions.
She bought her son a desk. Their diet included more protein and vegetables during a period when he was moving from sixth grade into junior high school. She left a high-pressure, unstable sales job for steadier hours. That gave them more shared dinners, homework time, and weekends together. The experiment later overlapped with a cancer diagnosis and an interruption in work, turning the payment into one of the few predictable parts of an otherwise unstable period.
The contrast with targeted welfare was revealing. Her application for low-income support took over two months, and compensation that included performance bonuses complicated the eligibility review. A system designed to identify need became another uncertainty imposed on a person already managing work, care, illness, and bureaucracy.
This is why people care about basic income even when the national debate sounds abstract. Cash can buy food and furniture, but its deeper effect is on time and the sequence of choices. It can let a parent change jobs before a crisis becomes unbearable. It can keep a health diagnosis from immediately becoming an income emergency. It can replace some of the mental calculation that consumes every day when a household has no margin for error.
The strongest evidence from a single-family experiment is not a claim that one story proves a national policy. It is a clearer view of what conventional indicators miss. Employment status does not tell us whether a parent can eat dinner with a child. Household income does not show how much uncertainty, delay, and humiliation the welfare system imposes. A payment can matter because it changes the problems a person has enough capacity to solve.
Su and Lee discuss AI access, economic security, the family experiment, child policy, and Taiwan’s route toward basic income on EP350. Image reproduced from the original BIEN report.
Child accounts reveal the difference between money now and ownership later
The episode’s discussion of the proposed 0-to-18 growth allowance and Taiwan Future Account was also more important than a debate over whether either policy deserves the UBI label.
A basic income has five defining features: it is periodic, cash, individual, universal, and unconditional. A payment limited by age is not universal across the whole population. A locked investment account is an asset rather than current income. A one-time deposit is not periodic. These distinctions should remain clear.
Still, Taiwan’s child-policy debate is assembling two parts of an income-and-ownership system.
A monthly child allowance addresses present liquidity. Families can use it for food, rent, school expenses, transport, or care. A future account builds capital that a young adult can use for education, a business, or another major transition. The policies answer different time horizons.
Taiwan should resist the temptation to choose between them. A child cannot eat compound interest, and a family facing this month’s bills should not have to wait until age eighteen for help. At the same time, current cash alone does little to correct the gulf in inherited assets that shapes adulthood. The stronger design combines accessible support now with protected capital later.
Lee connected this to low fertility in unusually human terms: “When survival pressure occupies all of a person’s mental energy, marriage and having children stop being life choices and become unreachable luxuries.”
That reaches beyond the standard claim that a subsidy might raise the birth rate. Young people care because Taiwan often treats marriage, housing, care, and children as private decisions while structuring the economy so that only households with family wealth can approach those decisions with confidence. Cash does not create a spouse, a home, or time. It can reduce the penalty for trying to build a life without inherited security.
The National Account proposal, under which government would seed investment accounts for children under twelve and the later Taiwan Future Account debate moved that idea into a larger interparty contest. This is how policy change often looks before it receives a coherent name: several parties approach the same insecurity through different instruments, each carrying part of the architecture.
The next step is already being prepared
EP350 is newsworthy partly because Ko’s involvement is continuing.
At the 2026 BIEN Congress in Toronto, UBI Taiwan and Ko’s team are planning a joint Taiwan forum. UBI Taiwan will present a youth basic-income policy framework focused on people aged eighteen to thirty-five. Ko’s team plans to present UBAI: AI Coin, treating access to capable AI as a public entitlement rather than a benefit tied to employment or the ability to pay for premium services.
Putting those proposals in the same session is intellectually useful. Youth basic income asks how people survive transitions. AI Coin asks whether they can access the tools shaping those transitions. The tension between them will expose the question most AI policy avoids: does the state want to make people more competitive inside the current distribution of ownership, or does it also want to change who receives the return?
Ko’s program is also preparing a further conversation with UBI Taiwan and JUJI, the fintech service operated within Gogolook, about a corporate-supported basic-income experiment. The emerging model would connect unconditional cash with a company’s financial-inclusion and social-impact strategy. That deserves a closer look. Corporate sponsorship is not a substitute for a public right, but it could test whether a pilot can generate operational knowledge, public attention, and a repeatable funding relationship instead of ending as a one-off charity project.
UBI Taiwan is meanwhile preparing a delegation for Toronto and working toward an Indo-Pacific basic-income summit in Taiwan for next year. The regional dimension matters. Taiwan’s route is emerging from semiconductor industrial policy, universal cash payments, child accounts, demographic pressure, and digital public infrastructure. South Korea has followed a different path through youth dividends, local currencies, and national electoral politics. Other Asian and Pacific societies may arrive through disaster payments, sovereign funds, rural guarantees, or public technology.
Asia does not need to wait for a Western government to produce a finished model and then copy it. Taiwan can contribute an approach rooted in the developmental state: the public helped build productive capacity, so the public should keep a claim on the return.
The political opening is real. It is also easy to waste.
There is a danger in treating this sequence as a victory. Universal Basic AI can become a voucher for technology companies. A child account can defer support until after the hardest years of caregiving. A sovereign wealth fund can accumulate assets for the state without ever creating an enforceable dividend for the public. A corporate pilot can produce a polished campaign and disappear when the sponsorship cycle ends.
Even the language of inevitability needs discipline. Lee said that he entered UBI Taiwan in 2020 expecting advocacy to be a lifetime cause because implementation seemed so distant. By 2026, he described himself as “hopelessly optimistic” and predicted that some form of UBI could emerge within five to ten years. He added that it may arrive under a different name.
I share the optimism about direction, but institutions do not become inevitable because the need is obvious. They become durable when advocates specify who receives money, how often, from which assets or revenue, through what legal claim, and with which protections against political capture.
The most plausible Taiwan pathway now has four connected parts.
A regular child or youth payment can establish an income floor for a politically visible group. A future account can establish a claim to capital. Universal Basic AI can establish access to high-quality digital capability. A professionally governed technology dividend fund can convert part of Taiwan’s publicly enabled AI wealth into cash paid to individuals.
The links among them matter; the labels are secondary. A future account without present cash leaves families squeezed. AI access without income leaves workers capable but dependent. A public fund without a dividend leaves citizens as spectators. Cash without an asset base remains vulnerable to the annual budget fight.
Taiwan has already shown that it can coordinate public investment, private enterprise, education, infrastructure, and long-term strategy to build an industry the world depends on. The political task is to apply the same seriousness to distribution.
The podcast is evidence of movement because the conversation now has somewhere to go. It leads back into the Legislative Yuan, forward to a joint forum in Toronto, outward to a corporate pilot, and toward a regional summit. Ko’s role is important because he can translate between technology policy and social policy at a moment when Taiwan can no longer afford to keep them separate.
The next generation’s oil field is already producing wealth. The unresolved question is whether people will receive tools to work around it, a savings account for the distant future, or an ownership claim they can feel in their lives now.
Taiwan should build all three and make the ownership claim permanent.
I am the founder of UBI Taiwan and participated in EP313 with Ko Ju-chun and Sarath Davala. I later interviewed Ko for my doctoral research and have been involved in UBI Taiwan’s discussions with his office. The interpretations and policy recommendations in this article are my own.
BIEN's first decade gave a scattered idea a name, a mailing list, and a place to meet
In September 1986, about sixty people came to Louvain-la-Neuve carrying a private suspicion: the idea they cared about had no home. They had arrived from different disciplines and political traditions, with interests running through philosophy, economics, social policy, labor, feminism, welfare reform, and the politics of unemployment. BIEN's own history calls the meeting a moment when many "lonely fighters" suddenly found others like them.
The Belgian beginning was small enough to name. In autumn 1983, Paul-Marie Boulanger, Philippe Defeyt, and Philippe Van Parijs, then connected with demography, economics, and philosophy at the University of Louvain, formed a working group around an idea Van Parijs had circulated in December 1982 under the phrase allocation universelle. The group used the collective pseudonym Collectif Charles Fourier. Its main public work was an April 1985 special issue of La Revue nouvelle.
The break came through a prize. The Collectif entered an essay competition on the future of work organized by the King Baudouin Foundation and won. The prize money did something ordinary and decisive: it paid for a meeting. Invitations and time had to be covered before an idea could become a network. The essay prize turned a circle of researchers into hosts.
The guest list now reads like a genealogy of the field: Gunnar Adler-Karlsson, Jan-Otto Andersson, Yoland Bresson, Hermione Parker, Annie Miller, Claus Offe, Guy Standing, Robert van der Veen, Georg Vobruba, Tony Walter, and many others. They did not arrive with a single doctrine. They arrived with overlapping worries about work, poverty, bureaucracy, freedom, and the design of European welfare states after the postwar settlement began to crack.
The proposal that held the room together was simple enough to say and difficult enough to argue over for forty years. Anyone who has been to a BIEN Congress knows we are still arguing. BIEN now defines basic income as “a periodic cash payment unconditionally delivered to all on an individual basis, without means-test or work requirement.” In the 1980s, that definition cut across the habits of social policy. Assistance usually meant proving need, proving household status, proving job search, or proving deservingness. Basic income moved the burden of proof away from the recipient.
The intellectual atmosphere was bolder than the later policy shorthand can suggest. Around the same time BIEN was being born, Robert van der Veen and Van Parijs advanced what they called a capitalist road to communism: a tax-financed unconditional basic income inside welfare capitalism, rising far enough to blur the line between paid work and free time. A 2024 revisit of that thesis connects it directly to automation and the post-labor world. The phrase still jars, which is part of its value. The early debate asked how much of life had been handed over to wages, and what might happen if survival loosened its grip on work.
The final session in Louvain-la-Neuve made the turn from conference to institution. Several participants wanted a permanent association that could publish a regular newsletter and organize regular conferences. Guy Standing proposed the name Basic Income European Network. The acronym mattered more than a committee might have expected: BIEN means good in French and Spanish. A serious organization began with a useful name and a small joke.
By 1988, the association had statutes. Its purpose was to connect individuals and groups interested in basic income and to foster informed discussion across Europe. Peter Ashby, Claus Offe, and Guy Standing became the first chairpersons. Walter Van Trier served as secretary, Alexander de Roo as treasurer, and Van Parijs edited the newsletter, later combining that role with the secretary position. The first decade depended on exactly this kind of administrative labor. Someone had to keep the addresses current. Someone had to collect papers, send notices, and make the next meeting happen.
The newsletter became the lifeline. BIEN's short history is blunt about the pre-internet reality: the printed newsletter formed the core of the network's existence. From 1988 to July 2001, it went to fee-paying members three times a year, thirty-six issues in all. In the first decade, the network traveled through envelopes. The movement had a pulse because paper kept arriving.
The congresses gave the pulse a calendar. Louvain-la-Neuve in 1986 led to Antwerp in 1988, Florence in 1990, Paris in 1992, London in 1994, and Vienna in 1996. The route is easy to flatten into a list of cities. It was more than that. Every stop carried papers, arguments, friendships, disagreements, and new people who could return home with proof that the conversation existed outside their own heads.
The anniversary responses add small human edges to the official chronology. In our interviews collected for the anniversary, Malcolm Torry remembered attending the Paris congress in 1992 and agreeing to help organize the London congress in 1994. Karl Widerquist remembered the same era from the outside: he was already a strong supporter of basic income and would have tried to attend BIEN events if he had known they existed, but in the pre-internet years he simply had not heard of them. That is the first decade in miniature: a network being built fast enough to matter, and slowly enough that people still missed it.
The public record also grew through books. BIEN's history lists proceedings from the first Louvain-la-Neuve conference, the Antwerp conference, Van Parijs's 1992 volume Arguing for Basic Income, and later collections that carried the debate into policy options, political feasibility, and income security as a right. The first decade was therefore an archive-making project as much as a meeting series. The papers did not disappear when the rooms emptied.
In 1996, BIEN opened a website, initially making its newsletters available to a wider public. Beginning in January 2000, the new BIEN NewsFlashes expanded the flow of online updates. That early digital presence mattered. For a network born by post, a public digital shelf changed the cost of discovery. Yannick Vanderborght later recalled buying the basicincome.org domain in the 2000s and learning to manage an early website; he remembered the design as terrible, but also said the digital presence helped expand BIEN's reach. The web did not create the network, but it made the network easier to find.
The first decade's achievement was quieter than legislation. BIEN gave the idea a name, a mailing list, a newsletter, a rhythm of congresses, a set of proceedings, and eventually a public address. By 1996, basic income supporters could point to something durable. They had a place to send papers, a place to meet, and a place to look for one another.
That made the later expansion possible when the network grew beyond Europe. The global turn did not come out of nowhere. It rested on the work of the first decade: the room, the mail, the congresses, the archive, and the stubborn habit of keeping the conversation alive.
A third round of NT$10,000 payments is moving through Taiwan’s legislature, towns are cutting their own checks, and one bill would make the handouts automatic. Taiwan keeps choosing cash, but it has not yet chosen to make the cash a permanent floor.
Taiwan is getting ready to hand out money again. A group of 23 KMT lawmakers, led by Lo Ming-tsai, has proposed another unconditional NT$10,000 payment, pitched as compensation for inflation from the Middle East conflict. The bill, with an estimated cost of NT$236 billion, cleared its first reading on May 8 and now sits in committee. At the same time, eight township governments from Penghu to Taitung were still paying out their own cash or subsidy programs in June, ranging from NT$2,000 to NT$10,000 per person. The 2025 national payment only closed its claim period on April 30. The country has barely finished one round of universal cash before opening the politics of the next.
This is the third reach for broad cash in three years. In 2023 the government sent every resident NT$6,000. In 2025 it sent NT$10,000, framed as a shield against United States tariffs and economic uncertainty. Now the opposition wants a 2026 repeat. Whatever else is true about Taiwanese politics, one habit has set in: when the economy wobbles or the budget swells, the reflex is to put cash in everyone’s hands.
For a basic-income audience, the shape of these payments is familiar. BIEN defines basic income as cash, paid to all, on an individual basis, without conditions or a means test. The handouts hit most of those marks. They are cash rather than vouchers, they are unconditional, and they reach essentially the whole population. The 2025 round even covered foreign permanent residents and the foreign spouses of citizens, which is broader than the household-registration basis expected for President Lai’s proposed child allowance (which would virtually be a child basic income). The element they lack is the one that matters most. They are not periodic. Each payment is a separate ad-hoc measure, justified by a separate issue, passed through a separate special act. A basic income is a floor people can count on. A handout is a windfall they hope arrives again.
What the townships reveal
The township programs are the most telling detail in the latest news, and the most revealing about where this leads. The eight still paying out in June ranged from post-disaster relief in flood-hit Guangfu and Renai to flat “economic revival” gifts elsewhere, from NT$2,000 in one township to NT$10,000 in another, and the tally announced across the year has climbed past a dozen localities. When local offices from Penghu to Changhua start cutting their own checks, the expectation of cash has spread well beyond the national level. That shows how normal the idea has become. It also shows the cost of doing it without a national floor.
The local version is uneven by design. Payments hinge on residency cutoff dates, run on whatever a township happens to have banked, and in at least one case have been delayed with no new date. Many are funded from one-off windfalls rather than stable revenue. In Chiayi County, ten township offices that struck it rich from sources like publicly run columbariums (cemeteries) are handing out cash while the indebted county government next door cannot, so neighbors in the same area get very different treatment depending on which office governs them. The money is real, but it spends down savings that could have gone to childcare, long-term care, or roads.
It also invites manipulation. Critics note that this is a local-election year, and in Hsinchu County one township’s residency cutoff drew more than a thousand people to transfer their household registration within days, shifting the balance of local council seats and prompting prosecutors to open an investigation. A single national payment, set at one rate and paid to everyone, would be simpler, fairer, and harder to game than a dozen towns improvising their own versions.
Why the cash keeps coming
It helps to be honest about why this keeps happening, because the mechanism says something about how basic income actually arrives. None of these payments came from UBI theory. The 2025 round grew out of a record tax surplus, a wave of recall votes, and a tariff shock. Critics accused the KMT of using universal cash partly to protect its own legislators, and the Cabinet asked the Constitutional Court to review it as unconstitutional. The payment went out anyway after recall campaigns against KMT legislators failed, because refusing to share a visible surplus had become politically impossible. Basic income, when it comes, is unlikely to arrive as a clean idea adopted on principle. It arrives as a crisis solved with cash, then repeated until people expect it.
The 2026 fight follows a similar script. Lo Ming-tsai justifies the new payment as relief from imported inflation, pointing to South Korea’s emergency oil-price support for its lower-earning households. The Executive Yuan opposes it. Premier Cho Jung-tai argues the surplus should go to debt and major projects and warns that another handout could feed inflation rather than ease it, and the finance minister urges fiscal discipline. The DPP caucus whip was blunter, saying his first thought was that an election must be near, and calling the bills pork-barrel.
The most consequential move is quieter. The KMT is also pushing a Budget Act amendment that would make cash handouts more automatic when annual tax revenue runs at least 15 percent above target and a surplus remains. If it passes, the handout moves closer to an automatic rule rather than an occasional act of will. The Executive Yuan opposes that too, arguing it would strip away the fiscal room a government needs in a real crisis. Both sides are now arguing, without quite saying so, about whether to make universal cash permanent.
There is a structural catch, and it points to the next argument worth having. Every one of these payments has been funded (or planned to) by a one-time surplus and justified by an emergency. That works for an occasional shock. It does not work as the basis for a permanent floor, because surpluses are not guaranteed and a crisis is not a budgeting plan. If Taiwan intends to keep returning national wealth to households, and three years of evidence suggests it does, it should fund that promise the way other resource-rich economies fund theirs, through a standing dividend rather than an annual scramble. Alaska pays its residents from an oil fund. Norway banks petroleum revenue in a sovereign wealth fund. Taiwan’s equivalent resource does not sit under the ground. It is the semiconductor and AI economy that already generates the surpluses these handouts spend. I have argued elsewhere for a Taiwan technology dividend built on exactly that logic.
This is also where the KMT’s surplus-triggered rule gets it backwards. Paying out only when tax revenue overshoots ties the public’s floor to the luckiest years and offers nothing in the lean ones, which is the opposite of what a floor is for. A dividend funded by a permanent endowment would pay in good years and bad. My earlier BIEN analysis described Taiwan as stumbling into an ad-hoc basic income driven by accumulated wealth and geopolitical anxiety rather than by ideology. The handout reflex shows the public is ready for recurring cash. The task is to fund it on purpose rather than improvise it from each year’s leftovers.
Two halves of the same trend
This is where the cash handouts and the child allowance belong in the same story. In a companion piece I argued that Lai’s proposed NT$5,000 monthly child payment matters because it would take Taiwan’s one-off cash habit and turn it into permanent architecture, at least for families with young children. The point that Henry Lee, a standing director of UBI Taiwan, made in that companion piece applies here as well: basic income will be realized through institutions, because people need it and governments need it, even when the name on the law is something else. The handouts and the allowance are two sides of the same movement. One shows that Taiwan will reach for universal cash again and again. The other shows what it looks like to make the cash regular. The task for basic-income advocates is to connect them, and to argue that the reflex behind the handouts and the structure behind the allowance should meet in a single, periodic, broadly universal floor.
The bottom line
Taiwan has now spent three years proving that universal cash is popular, workable, and politically irresistible. That is a real achievement, and basic-income advocates should say so plainly. The unfinished task is to stop treating each payment as a one-time favor. A handout that arrives once is relief. A payment that arrives on schedule, that people can plan around, that does not depend on the next surplus or the next election, is a basic income. Taiwan is closer to that than almost anywhere in Asia. It should decide, on purpose, to take the last step.
President Lai’s NT$5,000 monthly child allowance proposal is not a full UBI. But it may be the clearest basic-income stepping stone Taiwan has yet put into mainstream politics.
Editor’s note: This article describes an announced government plan moving toward legislation, not an enacted child UBI. Where it speaks of a “child basic income,” it means a stepping stone toward one rather than a finished program.
Taiwan’s basic-income debate has shifted. The government has not embraced full UBI, but in his second-anniversary address on May 20, President Lai Ching-te put something on the table that would have sounded politically unrealistic only a few years ago: NT$5,000 a month for every child from birth to age 18 as a new growth allowance. It is the centerpiece of an 18-measure population strategy that the Ministry of Health and Welfare casts as a move beyond subsidy-based childcare toward a continuous “0 to 18” system of family support.
The backdrop is a demographic emergency. Taiwan recorded about 213,000 births in 2015. In 2025 the number fell to 107,812, a record low, down about 20 percent in a single year and roughly half the total a decade earlier, and the country now has the world’s lowest birth rate. Cash for children is no longer a fringe proposal here. It has become the government’s headline answer to national decline.
It helps to see how Taiwan arrived here. For years the default was caution, built on vouchers, conditions, narrow welfare categories, and bureaucratic gates. Then in 2023 the island issued its first one-time universal cash payment of NT$6,000, and in 2025 it moved toward a second of NT$10,000 (and is now contemplating a third universal cash payment). Neither counted as basic income, since both were one-off rather than periodic, but they pulled universal cash into ordinary political practice. I wrote at the time that the second payment showed basic income gaining legitimacy as a policy direction in Taiwan.
The child allowance is a different kind of measure. Rather than a one-off tax refund, it is a recurring payment aimed at the whole of childhood. This implies an individual right to cash and, if enacted, will continue to push the overton window in terms of expanding this right beyond children. Lai put the annual cost at roughly NT$200 billion and argued Taiwan can afford it without crowding out other spending considering the growth of the economy due to the AI boom. The design splits in a way that should interest anyone who cares about basic income: children under 6 receive the full NT$5,000 in cash, while for ages 6 to 18 half goes to the family as cash and half into a locked “future account” the child can draw on at 18.
A child basic income stepping stone, with caveats
BIEN defines basic income as a periodic cash payment delivered to all, unconditionally, on an individual basis, without a means test or work requirement. A child allowance can be excellent policy and still fall well short of UBI.
By that standard, the allowance gets several things right. It is monthly, it spans the full 0–18 range, and preliminary planning does not exclude higher-income households. Its strongest piece, from a UBI perspective, is the 0–6 portion, which is unconditional cash stacked on top of existing child benefits.
Two caveats matter for a basic-income audience. First, the detailed eligibility rules have not been published, but officials expect the allowance to track Taiwanese household registration, as existing child benefits do. The payment would follow the child rather than the parents’ marital status, which is welcome, but a benefit keyed to household registration leaves out children who lack it. An allowance that is universal among registered children is not the same as one paid “to all,” and that gap is part of what separates a generous child benefit from a true basic income. Second, from age 6 the plan stops being pure income support, because half of each month’s money is locked away until adulthood.
That second feature sits at the heart of the debate. A monthly cash payment solves the problem a family faces this month, whether rent, food, tutoring, transport, or simply room to breathe. A locked account solves a different problem, asset inequality at adulthood. Both problems are real, but they are not the same problem, and money locked until 18 does nothing for a parent weighing today whether they can afford a second child.
Figure: Taiwan’s child allowance pays full cash for ages 0 to 6, then splits into half cash and half locked account for ages 6 to 18.
UBI Taiwan has pushed this for years
This proposal did not appear from nowhere. UBI Taiwan completed preliminary research in 2017 on implementing basic income in Taiwan and updated that policy-proposal work in 2018, including NT$5,000 a month for children. The child component has a history too, since monthly support for children was already part of the architecture in early discussions of a national UBI proposal. Our argument has stayed consistent: rather than patching holes after people fall through them, Taiwan should build a floor they can stand on before crisis hits.
UBI Taiwan’s single-parent pilot makes the case concrete. In one BIEN report, Ms. Yu, the program’s first participant, said the basic income gave her “a moment to breathe.” The pilot shows the same principle at the family level: when income is unstable, the freedom to decide where the money goes is what makes cash matter.
This proposal from the president is an important opening. The useful question is whether Taiwan is starting to normalize the institutional architecture of basic income under another name.
JiaKuan Su, chairperson of UBI Taiwan, responded to the proposal as an important step for Taiwan.
“I strongly affirm this policy. It breaks out of the traditional social welfare framework, responds precisely to the real needs of Taiwanese society today, and opens the possibility for a broader conversation about cash-transfer policy. However, while responding to the low-birthrate crisis, we also need to think about Taiwan’s long-term social foundation.
Only by comprehensively establishing basic economic security for individuals, and giving people real freedom, can we truly build the social resilience needed to face future challenges. This is the core spirit of UBI, and it is the indispensable next step for Taiwan to rebuild its social safety net,” Su said.
Henry Lee (Li Pinyi), a standing director of UBI Taiwan, said this was an important moment for the politics of basic income in Taiwan.
The spirit of the Lai administration’s policy fits the core elements of basic income, including being unconditional, cash-based, periodic, and paid to individuals. The only compromise is that the recipients are limited to parents of newborns and young children. This kind of policy is being framed as a necessary national investment and a response to the low-birthrate crisis, not as reckless cash giveaways.
Whether at street booths, on interview programs, or in exchanges at the Legislative Yuan, we are often asked: ‘How do you think UBI will be realized?’ Our consistent answer is that UBI will necessarily be realized through institutions, because people need it and governments need it. It is just that, when that day comes, its name may not be basic income.
Now, even though most Taiwanese have not yet noticed it, Taiwan is already moving toward an unconditional basic income system, as unstoppable as the development of AI.
UBI is unlikely to arrive in Taiwan as a finished theory dropped into the Legislative Yuan. It will arrive through the messy political process through institutions that solve concrete problems: child allowances, universal cash payments, youth security, AI-transition policy, and eventually a broader income floor. The task for basic-income advocates is to keep the design aligned with basic income principles, holding on to cash, regularity, individuality, universality, and trust.
The locked-account temptation
The opposition reached the cash-for-children debate first, with a rival model. In December 2025 the KMT and Taiwan People’s Party proposed a “Taiwan Future Account” for children 12 and under: an initial NT$50,000 deposit, NT$10,000 added each year until age 12, invested in Taiwan stock-index funds. Without parental top-ups, the account is projected to reach about NT$339,000 by age 18, or about NT$561,000 if parents add NT$10,000 a year. TPP chair Huang Kuo-chang framed it as a long-term investment in the next generation rather than a one-off subsidy. Focus Taiwan and CNA reported the design.
The partisan map is messier than ruling-party cash versus opposition accounts, though. The KMT also floated its own NT$5,000 monthly subsidy in May 2026, and Lai’s plan locks half the 6–18 payment into a future account that officials estimate will reach roughly NT$360,000 by 18, almost exactly the opposition’s figure. Once the branding is stripped away, both sides have converged on the same two instruments. The live question is not cash or accounts, but how much of each and at which ages. By that logic, the part of Lai’s plan that is a genuine basic-income stepping stone is mostly the full-cash 0–6 window.
Asset accounts do have real strengths. They give a child a visible ownership stake, they can teach long-term investing, and they are easy to sell politically as national investment rather than handouts. Through a basic-income lens, though, their limits are sharp. A locked account does almost nothing for next month’s rent, and nothing for the daily stress of young parents deciding whether they can afford another child now.
It also carries an equality problem. If richer parents can add top-ups and poorer ones cannot, an account that looks equal on paper drifts toward NT$561,000 for families with spare cash and NT$339,000 for those without. Taiwan’s health minister made the same point from the other direction, saying the government deliberately kept a personal-savings feature out of its growth allowance to avoid widening the gap between richer and poorer children, the very gap a top-up account invites.
Trump Accounts: good, but still not basic income
Taiwan’s account backers explicitly invoked America’s Trump Accounts, so it is worth being precise about what those are. The IRS and Treasury describe a $1,000 federal seed contribution for eligible U.S.-citizen children born from 2025 through 2028, created under the One Big Beautiful Bill Act signed July 4, 2025. By March 31, 2026, more than 4 million children had been enrolled, though only about 1 million were covered by the $1,000 pilot contribution. Left untouched, that seed is projected to grow to somewhere between $3,000 and $13,800 over 18 years.
Trump Accounts are one version of a child-asset-account tradition that runs through Cory Booker and Ayanna Pressley’s American Opportunity Accounts Act, the U.S. “baby bonds” proposal of a $1,000 birth account plus income-scaled deposits accessible at 18 for education or a home, and through the United Kingdom’s Child Trust Funds for children born between 2002 and 2011. These are capital grants. They may be useful, and they may be politically clever, but they are not periodic income.
The comparison actually flatters Taiwan. America’s federal child policy is a one-time $1,000 seed, while Taiwan’s is NT$5,000 every month, roughly NT$1.08 million per child over 18 years. If the model worth imitating were the generosity of recurring support, Taiwan is already ahead, which is precisely why it should not downgrade its own plan into a copy of a thinner American account.
What Taiwan should take, and what it should refuse
The lesson from Washington is not the branding. It is the political fact that governments increasingly accept that children deserve an asset floor, and voters are increasingly comfortable with universal benefits. Taiwan should take that and go further. The strongest design is cash plus accounts, not cash versus accounts, with the priority stated plainly: cash first.
In practice that means a few concrete choices. Keep the 0–6 payment as simple cash, unburdened by new conditions. If a future fund is used for ages 6 to 18, keep the cash half large enough to matter in real time. If parental top-ups are allowed, cap them or wall them off from the public guarantee, so the universal account does not quietly become another subsidy for families who can already save. And avoid means-testing at the point of receipt; if high-income households are the worry, recover the money through the tax system rather than forcing families through another application maze.
Universal cash is simple, visible, and trusted, while targeted welfare looks cheaper on a spreadsheet and turns expensive in bureaucracy, stigma, and exclusion errors. Taiwan’s 2023 and 2025 payments already proved the country can send broad cash and hold a serious public argument about shared prosperity. A monthly child allowance would turn that one-off habit into permanent architecture.
“When survival pressure fills up your mental bandwidth, having children stops being a life option and becomes an out-of-reach luxury.” — Henry Lee (Li Pinyi, 李品逸), UBI Taiwan standing director, writing in GVM
Lee’s warning reaches the core of it. A future account invests in the generation that arrives in 2044. The harder task is steadying the families who are deciding, in 2027, whether that generation arrives at all.
Why this matters for UBI in Taiwan
The proposal matters because it moves the battlefield. For years, basic-income advocates in Taiwan fought just to make universal cash sound respectable. Now the argument has moved to design: how universal the payment is, how much is locked away, and how permanent it becomes. That is a more productive fight to be having.
It also carries a warning. A government can adopt the language of basic income and then route most of the benefit into locked accounts, ending up with a policy that sounds visionary and fails the daily-life test. Families do not meet poverty only at 18. They meet it in monthly bills, unstable work, housing costs, childcare, and the constant low-grade tax of uncertainty. They also face unexpected crises that cash helps alleviate. Our basic income pilot recipient, Ms. Yu, was diagnosed with breast cancer in the middle of her basic income program and fortunately had this cash to help her navigate the instability associated with fighting such a disease.
Taiwan is a fitting place for the argument. It sits at the center of the AI and semiconductor economy, and that success has generated enormous value, yet many families still feel that raising children is financially irrational. Universal cash is one way to turn national wealth into household security.
A child basic income would not be the final destination for UBI in Taiwan. It would be a serious beginning, a statement that every child has a claim on the society they are born into and that every family deserves enough stability to make real choices. That is the moral center of basic income: a floor everyone can stand on, rather than charity routed through a maze of conditions.
Taiwan should welcome a child future account as a useful supplement for building assets. It should not mistake it for income security. The real breakthrough is the monthly cash, and the full-cash 0–6 window in particular. Get that right, and this could become one of the most important basic-income stepping stones in Asia.
UBI was already justified. AI is now creating its own independent reason to act faster.
Basic income was always justified because it is a foundation for human flourishing. Whether the economy is booming or falling does not change that principle. The rise of AI is compounding the justification because it will make an economy that provides vast wealth to some while alienating the majority.
Basic income advocates have always had to be careful with arguments related to automation. During the 2020 Democratic primary debates, I remember thinking candidate Andrew Yang was too early using automation as a justification for UBI. More importantly, framing basic income in this way distracts from the philosophical justification.
The strongest case for basic income never needed a prediction that paid work would vanish, and that caution was right.
But caution is not denial. At some point, refusing to update becomes its own kind of hype in the opposite direction. The serious question is no longer whether AI will instantly eliminate all work — it will not. It is whether AI is reorganizing the labor market faster than existing welfare states can handle. The answer is increasingly yes.
On June 4, Challenger, Gray & Christmas reported that U.S. employers announced 97,006 job cuts in May 2026, the highest May total since 2020. Employers cited artificial intelligence for 38,579 of them — 40 percent of all cuts that month, up from 7 percent in January. It was the highest monthly figure ever recorded for AI as a layoff reason since the firm began tracking it in 2023, and the third straight month AI led every other cause. For the year so far, AI has been cited in 87,714 cuts, already well past the 54,836 attributed to it in all of 2025.
That does not mean AI caused every one of those losses; companies have every incentive to use “AI” as a clean public explanation for ordinary cost-cutting. But the excuse is revealing. A firm does not blame layoffs on AI unless executives, investors, and workers already read AI as a restructuring force. Tool or pretext, management is now reorganizing work around it.
The same day, Anthropic published “When AI builds itself,” a warning about recursive self-improvement. The company said it is delegating a growing share of AI development to AI systems themselves; that its engineers now ship roughly eight times as much code per quarter as they did from 2021 to 2025; and that more than 80 percent of the code merged into its codebase is now attributable to Claude. It stressed that recursive self-improvement is not here and not inevitable but warned it could arrive sooner than most institutions are prepared for.
Set the two side by side. At one end of the economy, managers are using AI to justify record layoffs; at the frontier, engineers are using AI to speed up the building of AI itself. Neither report causes the other, but together they widen the range of futures a welfare state has to plan for — and that changes the policy calendar.
The next morning, June 5, the Bureau of Labor Statistics reported that the broad labor market was holding: total nonfarm payrolls rose by 172,000 in May, and unemployment was unchanged at 4.3 percent. Anyone claiming AI has already produced mass unemployment is overstating the evidence. Even Andy Challenger, whose own firm tallied the record AI layoffs, called it “not yet the jobpocalypse some predicted.”
But waiting for economy-wide unemployment to spike before acting would be a policy failure. AI disruption may not arrive first as breadline unemployment. It may arrive as fewer entry-level jobs, slower hiring into exposed occupations, reduced hours, compressed wages, contractor instability, and a quiet shift in bargaining power from workers to firms.
Anthropic’s own labor-market research is cautious on exactly this point. It found no clear unemployment effect yet for the most AI-exposed occupations, but tentative evidence that workers aged 22 to 25 are entering those professions at lower rates. That is the kind of early warning policymakers usually miss (or ignore), because it looks less like a crash and more like a missing rung on the career ladder.
The old welfare state was built to respond after a worker loses a job. AI may weaken people before that moment: fewer interviews, lower freelance rates, fewer junior openings, more unpaid retraining, more pressure to accept whatever terms remain. If the harm is diffuse, delayed, and hidden inside “productivity,” unemployment insurance alone will not catch it.
The public conversation keeps getting distracted by the wrong question: will AI become humanlike? That is not the threshold that matters for workers. AI does not need consciousness, emotions, or science-fiction personhood to change the economy. It only needs to perform enough economically valuable tasks cheaply enough that companies reorganize around it.
That is already happening. Anthropic’s Economic Index found that about 49 percent of jobs have seen at least a quarter of their tasks performed using Claude, up from 36 percent in early 2025 — and the work is migrating from casual chat interfaces toward API and agentic workflows that look more like automation than ordinary assistance.
The first AI labor shock is rarely a robot replacing a whole worker. It is a manager asking why a team of ten cannot become a team of six, a company deciding not to hire juniors because senior staff with AI tools can stretch further, and the slow collapse of the pipeline that used to turn beginners into experts. The worker stays employed but now competes against a machine that never asks for health insurance, rent, sleep, or dignity.
Even if recursive self-improvement is not imminent, the direction is clear: AI is increasingly used to build, test, debug, and accelerate AI. That creates a compounding force. If AI helps build better AI, the pace of change becomes less tied to human institutional speed and more tied to compute, capital, data, energy, and competition.
Anthropic itself sketched a future in which humans play a diminished role in AI development, shifting toward oversight and verification of a growing virtual lab run by AI systems — and admitted it is hard to predict what the economy looks like if human labor stops being competitive. A leading AI company is effectively saying it is not sure whether the economy can absorb what it is building, or whether institutions are moving fast enough. Then build the floor now.
The strongest basic income argument remains human, not technological. People need security because life is unstable, and bargaining power because employers, landlords, creditors, and bureaucracies often hold too much control over daily survival. They need room to leave abusive relationships, care for children, recover from illness, study, relocate, or simply breathe. That was true before ChatGPT and would remain true if the boom stalled tomorrow.
What AI changes is the risk profile. Conditional welfare asks the state to decide who is poor enough, unemployed enough, disabled enough, or deserving enough — and AI disruption will not respect those categories. A designer may lose contracts without ever appearing unemployed; a junior lawyer may never be hired in the first place; a teacher may keep the job but be expected to produce twice as much for the same pay. A universal cash floor is not perfect, but it is built for uncertainty: it reaches people before the form of harm is legible to the bureaucracy.
Reskilling matters, but it is not enough. The old answer to technological change was education: learn the new tool, move up the value chain, become more adaptable. That is useful advice at the individual level, but it is not a social contract.
The problem is that AI is climbing the skill ladder with us. When automation threatened factory work, the advice was to move into knowledge work; when AI threatens knowledge work, the advice becomes to supervise AI. But not everyone can become an AI manager, and even AI managers can be made productive enough that fewer are needed. A society cannot build economic security on an endless game of musical chairs in which every worker is told to outrun the next model release. Basic income does not replace education — it makes education realistic, giving people time to retrain without panic, reject bad jobs, and take risks without falling through the floor.
If reskilling alone cannot carry the weight, what can? The basic income world is already testing one answer. In March 2026, BIEN reported on an “AI Dividend” run by two nonprofits, the AI Commons Project and What We Will, sending $1,000 a month for a year to a cohort of 25 to 50 workers who had lost pay, jobs, or opportunities to AI, on about $300,000 in initial funding with hopes of distributing $3 million in 2026.
The revealing part is how its organizers plan to scale: by lobbying profitable AI companies to donate. That is the right instinct — the wealth AI creates should reach the people it disrupts — wired the wrong way, as charity that depends on the goodwill of the firms doing the disrupting. This is where the debate should go next. Not sympathy for a handful of displaced workers, and not a temporary patch for those who can prove AI harmed them, but a standing public claim on AI-era wealth, paid out in cash. The right language is joint ownership instead of simply compensation.
AI is not built out of nothing. It rests on public research, public universities, open scientific knowledge, government procurement, legal systems, energy grids, water systems, semiconductor supply chains, publicly educated workers, and a vast archive of human culture. Then the gains concentrate: model companies, cloud and chip firms, data-center owners, and major shareholders capture the first wave of wealth, while workers are told that disruption is the price of progress and families are told to be flexible.
If AI creates a productivity windfall, the public should receive part of it automatically. If AI creates labor disruption, the public should receive protection automatically. And if AI companies depend on public infrastructure and public knowledge, the public should hold a permanent claim on the upside. The cleanest version of that claim is an AI Social Dividend Fund. Bernie Sanders’ proposed 50 percent one-time tax on AI stocks is ethically justified, this is likely to be too economically distorting to maximize the upside for the public.
A practical AI Social Dividend Fund
The design should be simple enough to explain and hard enough to corrupt.
Public upside for public support. When frontier AI firms receive major public support — subsidies, procurement contracts, energy priority, infrastructure buildout, tax incentives, government-backed financing, or strategic regulatory privileges — the public should receive warrants, equity, revenue participation, or another upside instrument. Predictable and contractual, not a surprise penalty: if the public shares the risk, it shares the return.
Real prices for real constraints. Compute and infrastructure fees should reflect the scarce electricity, land, water, and grid capacity AI consumes. Pricing those bottlenecks is basic economics, and a portion of the revenue should flow into the fund rather than disappearing into general budgets.
Insulated, professional management. The fund should be globally diversified, transparent, and legally protected from election-year raids — not a slush fund for industrial policy. The point is to convert AI-era rents into public wealth, then convert part of the returns into a cash dividend.
A simple, universal cash payout like the Alaska Permanent Fund. The dividend should go to individuals in cash, regularly, universally, and without work conditions. High earners can be taxed back later through the income-tax system; do not build a humiliating bureaucracy at the point of receipt. The payment should be simple because the future will not be.
Why universal beats targeted in the AI era
Targeting sounds efficient until the world changes faster than the eligibility rules. Who counts as “AI-displaced”? The worker laid off after a company adopts AI? The freelancer whose rates collapse? The graduate who never gets the first job? The employee whose workload doubles while pay stays flat? The artist whose style is absorbed into a model but whose loss never shows up in unemployment statistics? A targeted relief system would spend years fighting over proof. A universal cash floor skips the impossible detective work: the same transition that raises national wealth should raise household security.
Universalism also protects politics. Programs only for the visibly displaced become narrow, stigmatized, and easy to cut. A universal dividend builds a broad constituency that reads it as a share of public wealth, not a handout for failure.
The worst criticism of UBI in the AI debate is that it gives up on work. It gets things backwards. Basic income is how people keep agency when work changes. It is a platform for caring for others and taking risks.
In an AI economy this matters even more. If AI raises productivity, people should gain more freedom. If it compresses some jobs, people should gain more mobility. If it creates vast fortunes, the public that made it possible should gain more ownership — not just more advice to adapt.
The temptation will be to reach for the old, comfortable failures: paternalistic vouchers that tell people the state knows their lives better than they do; job training offered with no income behind it, which is just homework assigned during a crisis; or a narrow “AI-victim” bureaucracy that arrives too late and forces workers to prove harm in a market where harm often looks like an opportunity that never came. Unemployment insurance belongs on that list too — it was built for a clearer employment relationship than the AI economy is likely to provide.
And the case itself should not be rooted in the most extreme AGI scenario. That makes the policy look speculative when the real evidence is already visible in layoffs, hiring shifts, wage pressure, and concentrated ownership. We do not need a forensic unit to trace AI disruption back to its source. We need a universal floor.
The politics of AI and basic income are not as far apart as they look. The left can support an AI dividend because it redistributes concentrated technological gains and protects workers. The center can support it because it is simpler than expanding dozens of targeted programs that always lag reality. Conservatives can support it because it strengthens families, rewards public ownership, avoids paternalistic bureaucracy, and hands people cash rather than another managed-service maze.
The national-security case is stronger than it first appears. Societies under rapid technological stress become angry, distrustful, and vulnerable to demagogues. If people experience AI as a machine that enriches a few and destabilizes everyone else, they will turn against both the technology and the institutions that allowed it.
The bottom line
Basic income did not need AI to become morally serious. But AI has made the timeline harder to ignore. The responsible position is not panic; it is preparation. The labor market has not collapsed, yet employers are already citing AI for record job cuts. Recursive self-improvement has not arrived, yet a leading lab now says AI is increasingly building AI and that institutions may not be ready. The evidence does not show that the future is fixed. It shows that the range of plausible futures has widened, and the downside risks are now large enough to demand architecture.
A society that waits for perfect proof of disruption will build its support too late. By the time every dataset agrees, the bargaining power will have shifted, the entry-level jobs will have thinned, the wealth will have concentrated, and the public will again be told to accept insecurity as the cost of progress.
The better answer is direct and practical: build a universal cash floor, funded in part by the AI-era wealth that public knowledge and public infrastructure helped create.
AI may or may not become recursively self-improving soon, and the economy may or may not face mass technological unemployment. But the policy lesson is already clear.
If machines are going to learn faster, society has to protect people sooner.
Note on sources
Linked inline: BIEN’s basic income definition and its March 2026 report on the AI Dividend; the Challenger, Gray & Christmas May 2026 job-cuts report; Anthropic’s “When AI builds itself,” Economic Index, and labor-market-impacts research; Scientific American on the recursive-self-improvement warning; the U.S. Bureau of Labor Statistics May 2026 employment situation; Jones & Marinescu on the Alaska Permanent Fund Dividend (AEJ: Economic Policy, 2022); and the Stockton Economic Empowerment Demonstration.
