Losing a job always hurts, but losing one to a machine tends to hurt for longer. A new note from Goldman Sachs suggests that the first Americans to be pushed out by artificial intelligence are already following the same rough path that millions of workers walked during earlier waves of automation.
The timing is hard to ignore. Plenty of employers have pointed to AI this year when trimming staff. Visa was among the latest in late July, when it moved to cut roughly 7% of its global workforce, or about 2,600 jobs. For the workers who receive those notices, the real question is how long the road back will take.
Goldman sees history repeating for AI-displaced workers
Goldman economist Pierfrancesco Mei found that workers displaced from AI-exposed jobs since 2025 needed roughly a month longer than other laid-off workers to land a new role, Fortune reported. They were also about 10 percentage points more likely to still be out of work a year after losing their jobs. A gap Mei said matched what the bank saw in past rounds of technological change.
Also read: Open AI’s Sam Altman has a chilling warning for young job seekers
This isn’t Goldman’s first look at the problem. In April 2026, Mei wrote that workers who lost jobs to technology took about a month longer to get hired again. And that they earned more than 3% less in real terms when they did.
He blamed what he called occupational downgrading. Many of those workers ended up in roles that required fewer of their analytical and people skills.
The newest findings rely on the Displaced Worker Supplement, a survey the government has run every two years since 1984 alongside the Current Population Survey, according to the Bureau of Labor Statistics. It tracks what happens to people after an involuntary layoff.
Since the Displaced Worker Supplement does not ask whether AI cost someone a job, Goldman used a composite measure of AI exposure to identify likely displaced workers. The bank found no meaningful gap in outcomes for job losses before 2022, when broad AI adoption began to take hold.
Older workers carry the heaviest burden
Age matters a lot here. Laid-off workers 35 and older from AI-exposed jobs needed about five more weeks than other displaced workers to get back into full-time work. For those under 35, the delay was closer to three weeks. That also fits the old story, since older workers also had the roughest time in earlier tech shakeups.
The government’s broader numbers, which cover all kinds of layoffs, look much the same. Some 7.4 million workers were displaced between January 2023 and December 2025, up from 6.3 million in the previous survey. Among people who had held their jobs for at least three years, those 55 and older were far less likely to be working again by January 2026 than workers ages 25 to 54, as reported by the Bureau of Labor Statistics.
There are signs that AI is nudging older professionals out the door too. Economist Geoffrey Sanzenbacher of the Center for Retirement Research at Boston College found in June 2026 that older workers in highly AI-exposed occupations have been leaving their jobs at a faster clip since ChatGPT arrived in late 2022. A good share of those exits ended in unemployment, not retirement, according to the Center for Retirement Research.
Even those who find work again are often taking a pay cut. About 49% of long-tenured workers who were rehired into full-time jobs earned as much as or more than they did before, down from about 62% in the previous survey.
Goldman economist Pierfrancesco Mei found that workers displaced from AI-exposed jobs since 2025 needed roughly a month longer than other laid-off workers to land a new role.
Young workers face a different kind of squeeze
Younger workers who do get laid off are faring better, at least for now. Goldman found that they took fewer temporary jobs on the way back to full-time work and were less likely to supplement their income with a second job. Mei flagged that as a relief, given how much concern has built up around AI’s threat to new graduates.
The bigger problem for young people may arrive before they are ever hired. Stanford researchers using ADP payroll data found in August that employment for workers ages 22 to 25 in the most AI-exposed occupations sits about 19% below where it would have been had it kept pace with peers in less-exposed jobs. The gap appears to stem mostly from less hiring, not layoffs.
More Goldman Sachs:
- Goldman Sachs rethinks U.S. economy as interest rates stay higher
- Goldman Sachs spills the beans on what’s next for S&P 500
- Goldman Sachs sees big change coming for interest rates
Harvard researchers found something similar inside companies. Firms that adopted generative AI cut junior headcount by about 7.7% within six quarters, while senior staffing held steady. And in a June survey by GMAC, a third of employers said they had used AI to handle some entry-level work instead of hiring a person, as reported by TheStreet.
Students appear to be adjusting their plans. Goldman found that college students are increasingly choosing majors that lead to jobs with lower AI exposure and stronger growth, though the bank cautioned that it is still too early to judge the full impact.
Economists remain split on how deep the damage runs
Employers are certainly talking about AI more than ever. Companies cited it in 120,136 announced job cuts through September, roughly 21% of all cuts and the top reason given this year, though it made up only about 9% of cuts in September itself, according to Challenger, Gray & Christmas. Tech firms led all industries. In the spring, the number of AI-cited cuts hit the highest monthly level since Challenger started counting in 2023.
Not everyone is convinced AI is rewriting the labor market at scale. The Yale Budget Lab looked at employment patterns tied to AI exposure and found little deviation from earlier trends. Goldman’s own data has a blind spot, too: it only captures workers who held a job before losing one, which means some of AI’s impact may show up in roles that simply never get filled.
Related: Goldman Sachs has a new warning for the housing market for 2027
