Economists Sound the Alarm on AI Jobs as Young Workers Feel It First

More than 200 economists, including 16 Nobel Prize winners, signed an open letter Monday warning that artificial intelligence (AI) could transform the economy faster than the Industrial Revolution. Organized by Stanford University’s Digital Economy Lab, the letter runs just four sentences and carries signatures from computer scientists and executives at Anthropic, Google and OpenAI.

The letter speaks mostly in the conditional tense, warning what AI “could” do to jobs and living standards. One slice of the workforce isn’t waiting to find out. Recent college graduates are already posting unemployment numbers well above the national average, even as the broader economy looks calm on paper.

More Than 200 Economists Sign a Four-Sentence Warning

The statement is titled “We Must Act Now” and packs its warning into four sentences with no attached policy plan.

“AI may become radically more powerful over the next 10 years,” the letter reads. “This could drive an unprecedented transformation of our economy, larger than the Industrial Revolution, but unfolding over a vastly shorter time frame. It could bring risks, including large-scale job displacement, as well as opportunities such as major gains in living standards.”

The letter calls on leaders to build “the incentives, guardrails, and institutions needed to steer AI in a direction that complements humans and benefits society.” Signatures span top economists, computer scientists and sitting executives at Anthropic, Google and OpenAI, the very companies building the technology in question, alongside a run of Nobel laureates rarely seen on a single document.

Computer scientist Yoshua Bengio, a University of Montreal professor and one of AI’s founding researchers, signed on and went further in a separate statement, saying that based on the technology’s trajectory, “it is highly plausible that AI will drastically transform our economies.”

We must be intentional and make collective, democratic choices, rather than letting market forces play out and risking leaving most citizens behind.

Bengio wrote that line in the same statement, arguing the shift is too consequential to leave to chance.

The Layoffs Companies Already Blame on AI

Corporate layoff trackers already show what the letter only forecasts. Challenger, Gray & Christmas, the outplacement firm that has tracked job cuts for decades, counted nearly 102,000 layoffs attributed to AI so far this year, with the tech sector alone responsible for a third of all 2026 job cuts.

“It’s certainly making an impact as we speak in a way that no technology has before,” said John Challenger, the firm’s chief executive. He expects the disruption to spread further. “Finance might be the next big sector that’s most affected,” he said.

The pattern isn’t confined to one sector. A California Policy Lab tracker found finance and insurance had the state’s highest concentration of unemployment claims coming from workers in highly AI-exposed occupations, with information and professional services close behind. Researchers there said the numbers suggest AI’s effects “may be starting to surface,” even though statewide claims don’t yet show a broad wave of AI-driven job loss.

Pooja Sriram, senior U.S. economist at Barclays, isn’t ready to blame AI alone. “Some of this could genuinely be productivity replacing workers,” she said.

A Stanford Digital Economy Lab study found employment weakening in tasks AI automates outright, while it held up in jobs where the technology mostly augments what workers already do. The Bureau of Labor Statistics now projects office and administrative support roles, jobs that make up roughly a quarter of employment in financial services, will see some of the steepest declines of any category over the next decade.

Why Are Recent Graduates Struggling to Find Work?

Recent college graduates face unemployment well above the national rate, and researchers increasingly tie the gap to employers leaning on AI for entry-level tasks like drafting, research and basic coding. A Federal Reserve Bank of New York analysis put joblessness among workers aged 22 to 27 at 5.6% at the end of last year, against 4.2% for the workforce overall.

Yale School of Management researchers, drawing on National Bureau of Economic Research data, describe a similar gap widening even faster: recent-graduate unemployment has climbed to nearly 6%, rising twice as fast as the rest of the workforce since 2022. Yet the same body of research found little to no measurable impact on employment or productivity in almost 90% of firms surveyed, based on responses from nearly 6,000 executives.

The job market looks calm in the aggregate, and looks different once the youngest, least senior group is isolated. Researchers point to a separate, longer-running trend compounding the squeeze: the earnings gap between workers under 35 and those over 55 has widened by more than 60% over the past four decades, as older employees hold on to roles longer and slow the career ladder for everyone behind them.

Where Economists Split on the Timeline

Signing the same letter doesn’t mean agreeing on the speed of the threat. Even among researchers who study AI’s labor effects for a living, the estimates range from a handful of years to a full decade, and from modest disruption to mass displacement.

• Daron Acemoglu, the Nobel-winning economist, says he hasn’t abandoned his doubts that AI will move as fast as the industry’s most optimistic voices claim, even as recent breakthroughs have sharpened his worry about workers losing jobs.
• Dario Amodei, Anthropic’s chief executive, has said AI could displace half of entry-level white-collar jobs within five years, and would “bet pretty strongly against” job creation fully offsetting the losses.
• Erik Brynjolfsson, the Stanford economist who helped organize the letter, told The New York Times there has been “a notable change in the profession,” and says the most urgent task now is fixing years of contradictory data on who AI is actually displacing.

Anthropic itself has started building tools to answer that measurement problem. The company published a new framework tracking unemployment among AI-exposed occupations, arguing that job postings alone can’t show whether displaced workers are actually finding new work elsewhere.

An Industrial Revolution, Compressed

The letter’s own comparison sets a high bar. The Industrial Revolution took the better part of a century to move economies from farm to factory. Signatories argue AI could compress a shift of similar size into a fraction of that time.

History offers a reason for caution on both sides. More than 60% of jobs in 2018 were in professions that did not exist in 1940, a reminder that past waves of automation destroyed and created work in roughly equal measure. Anton Korinek, a University of Virginia economist currently embedded with Anthropic, framed the stakes in the same historical terms. “Steam, electricity, and computers each gave societies decades to adapt,” he said.

A separate survey of academic economists, frontier AI company staff, policy researchers and elite forecasters, run by researchers at the Federal Reserve Bank of Chicago, Yale, Stanford and the University of Pennsylvania, found a more measured consensus: every transformative general-purpose technology in modern history, from electrification to the personal computer, took decades to show up in aggregate productivity statistics. AI, in most economists’ telling, is unlikely to be different.

States Move Faster Than Washington

Washington’s response so far is a research hub with no accompanying rulebook. The Trump administration’s 2025 AI Action Plan directed the Department of Labor to stand up an AI Workforce Research Hub to gather better labor data, while the Census Bureau still tracks adoption through a survey that asks companies only whether they use AI, not how much or where.

States have moved with more specifics. A politically mixed group of them has adopted some version of the human-in-the-loop automation law California passed in 2024, alongside newer experiments with a right to retraining for displaced workers.

• Human-in-the-loop rules – require a person to review automated decisions before they take effect, first adopted in California.
• Right-to-retraining provisions – give displaced workers a claim on employer-funded reskilling.
• Sector-specific brakes – slow automation in health care, education and creative production, where consumer-safety concerns overlap with job worries.

Brookings researchers, tracking these state-level guardrails, describe skilling programs as the most politically popular response and the least sufficient one on its own, since training does not guarantee the good jobs displaced workers are actually looking for.

That compressed timeline was the subject of a Hoover Institution panel in March, when Hoover Institution Director Condoleezza Rice recalled asking Fei-Fei Li, a Stanford computer science professor and AI pioneer, what was coming next.

“You mean in six weeks?” Li replied, Rice recounted at the same Hoover Institution panel. Former U.K. Prime Minister Rishi Sunak, on the same stage, put the politics in blunter terms: AI, he said, is “less popular than almost anything else that you can think of, including politicians from all parts.”