Published 23 October 2026
The $25-an-hour question: is a job guarantee the answer to AI mass unemployment by 2028?
In 2026, the predictions read like science fiction. Bill Gates said humans would not be needed “for most things” within a decade. Anthropic CEO Dario Amodei warned that AI could wipe out swathes of law, tech, finance and white-collar jobs within one to five years. A fictional memo from a future June 2028, penned by investment firm Citrini, described an American economy where unemployment had exceeded 10 percent, the S&P 500 had crashed 38 percent, and high-earning professionals had been forced into gig-economy poverty.
By 2027 and 2028, these scenarios are no longer hypothetical. They are the competing narratives shaping a furious policy debate: if AI does displace millions, should the government simply guarantee everyone a job?
The case FOR a job guarantee
It protects more than a paycheck. A job guarantee does something universal basic income cannot: it preserves the dignity, routine and social connection that work provides. Venture capitalist Nick Hanauer put it plainly: “Jobs represent so much more than income in human societies. They represent connection and dignity and enfranchisement and participation.” Algernon Austin of the Center for Economic and Policy Research agreed: “When you have a job, you gain work experience and you gain skills and you gain a network.” A UBI cheque cannot replace any of that.
It pays for itself, sort of. Oxford economist Max Kasy, who studied a recent Austrian job guarantee experiment, argues that the math works when you count the savings. “The cost saving comes from the fact that you don’t have to pay unemployment benefits, or maybe disability benefits or early retirement benefits. Then people who are working pay taxes and make social insurance contributions, so that’s additional revenue for the state.” Pavlina Tcherneva of Bard College goes further: “It’s not only affordable, it’s an investment.” She points to the healthcare costs generated by long-term precarious unemployment, costs that disappear when people are in meaningful work.
It tackles real problems while it protects workers. A job guarantee would not simply pay people to sit idle. The jobs would include climate adaptation, elder care, childcare, disaster recovery and public infrastructure. In the 1930s, FDR’s Works Progress Administration built roads, bridges, sewers and playing fields while employing millions. A modern version could retrofit buildings for energy efficiency, expand aged-care capacity and harden communities against floods and fires, all while giving displaced professionals a bridge to the next chapter of their careers.
It is already gaining political traction. US Representative Ayanna Pressley has reintroduced a federal job guarantee resolution calling for at least $25 an hour, adjusted for cost of living. A separate proposal backed by Representatives Greg Casar, Valerie Foushee and Sara Jacobs would tax AI companies on their token usage or AI revenue specifically to fund the guarantee. If unemployment rises, the tax rate rises. In Cleveland, the non-profit Universal Basic Employment is already running a pilot that splits the cost of a $27 “autonomous wage” with employers, giving workers a raise without pricing them out of a job.
It is more popular than UBI. Global polling consistently shows stronger public support for job guarantees than for universal basic income. David Shor’s polling of Americans shows the same pattern: people prefer the creation of well-paid jobs to direct cash handouts. Devin Cotten, the Cleveland pilot’s founder, says the political argument is simple: “Work is a shared value across political lines. I think crafting a solution based on work allows us to be a bipartisan solution.”
The case AGAINST a job guarantee
The unemployment apocalypse may be overstated. Goldman Sachs Research, one of the most cited voices on AI and jobs, projects that AI will displace roughly 6 to 9 percent of workers over a ten-year adoption cycle, not 50 percent. Even at the high end, the bank expects the unemployment rate to rise by less than one percentage point in any given year. Harvard economist Doug Elmendorf, former director of the Congressional Budget Office, co-authored an NBER paper with four scenarios. His cautious conclusion: “I think it’s very likely that there will be significant job losses from artificial intelligence in the coming couple of decades, but that most, not all, but most of those people will be able to find new work.”
History suggests technology creates more than it destroys. The World Economic Forum’s 2025 Future of Jobs Report projects 92 million jobs displaced by 2030, but also 170 million new ones created. That is a net gain of 78 million positions. Yes, displacement is front-loaded in 2027-2028, and yes, the new jobs demand skills many displaced workers lack. But the lesson of previous industrial revolutions is not that work disappears, but that work transforms.
The fiscal reality is brutal. A federal job guarantee paying $25 an hour to even a modest fraction of displaced workers would cost hundreds of billions annually. During a US Senate hearing, Republican Senator John Kennedy told Century Foundation president Julie Margetta Morgan: “Do you believe in the Easter bunny?” Critics argue the money is better spent on targeted retraining, modernised unemployment insurance and portable benefits, interventions that help workers pivot rather than park them in a public-sector queue.
It risks creating a permanent underclass of public-sector workers. There is a fine line between a job guarantee and workfare. If the government becomes the employer of last resort for millions, the jobs could become low-wage, low-prestige dead ends with no career ladder. The program could also depress wages in the private sector by setting a wage floor that small businesses cannot match, or by flooding local labour markets when private-sector hiring eventually recovers.
AI is augmenting jobs as fast as it replaces them. A 2025 McKinsey survey found that while 32 percent of organisations using AI expected job cuts, only 14 percent had actually seen them. Accenture reports that 52 percent of top executives expect AI to increase entry-level hiring, not decrease it. Gartner analysed hundreds of large organisations and more than one million workers: AI-driven layoffs accounted for less than 1 percent of tracked workforce reduction. Most job losses still come from shifting market demand, mergers and macroeconomic conditions, not from artificial intelligence.
The verdict
The honest answer is that nobody knows whether 2027 and 2028 will bring mass unemployment or a manageable transition. The range of credible estimates runs from Goldman Sachs’ modest displacement, less than one percentage point on the unemployment rate, to Citrini’s fictional 10.2 percent unemployment and 38 percent market crash.
What is clear is that the costs of being wrong are asymmetrical. If policymakers under-prepare and a crisis hits, the ideas lying around better be good ones. A job guarantee is not a silver bullet. It would be expensive, administratively complex and politically divisive. But it also offers something no other policy can match: a direct, dignified bridge between the jobs AI takes and the jobs society still needs done. As Friedman observed, crises create change. The question is whether we have the courage to refine the right ideas before the crisis arrives.
How this analysis is researched, sourced and corrected: Editorial standards and corrections. Spotted an error? Contact Philip Hall.
