Losing a job is always painful, but losing one to a machine often causes longer pain. A new report from Goldman Sachs suggests that the first Americans to be displaced by artificial intelligence are already following the same rocky path that millions of workers traveled in earlier waves of automation.
The timing is hard to ignore. Many employers have mentioned artificial intelligence when laying off workers this year. Visa was one of the latest moves in late July, when the company cut about 7% of its workforce, or about 2,600 jobs, globally. The real question for workers receiving these notices is how long the road back will take.
Goldman Sachs thinks history will repeat itself as workers replaced by artificial intelligence
According to Fortune magazine, Goldman Sachs economist Pierfrancesco Mei found that since 2025, workers displaced from jobs affected by artificial intelligence will take about a month longer than other laid-off workers to find new jobs. They are also about 10 percentage points more likely to remain unemployed one year after being unemployed. The gaps May said are consistent with what the bank has seen in past rounds of technology change.
Also read: Open AI’s Sam Altman issues chilling warning to young job seekers
This is not the first time Goldman Sachs has focused on this issue. In April 2026, May wrote that it would take about a month for workers displaced by technology to be hired again. And in real terms, their income will be reduced by more than 3%.
He blames this on so-called career demotion. Many of these employees end up in positions that require less analytical and interpersonal skills.
The latest findings rely on the Unemployed Workers Supplement, a survey the government has conducted every two years since 1984 in conjunction with the Current Population Survey, according to the U.S. Bureau of Labor Statistics. It tracks what happens to people after they are involuntarily laid off.
Because the Displaced Workers Supplement does not ask whether AI will cost someone their job, Goldman Sachs used a composite measure of AI exposure to identify workers who might lose their jobs. The bank found that there will be no significant difference in unemployment outcomes until 2022, when AI begins to be widely adopted.
Older workers bear the greatest burden
Age matters here. Workers aged 35 and older who were laid off due to AI took about five weeks longer than other unemployed workers to return to full-time employment. For those under 35, the delay is closer to three weeks. This also fits the age-old story that older workers also had the hardest time during early technological changes.
The government’s broader figures, covering a variety of job cuts, look broadly 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 have been working for at least three years, those 55 and older are far less likely to be working again by January 2026 than workers ages 25 to 54, according to the Bureau of Labor Statistics report.
There are signs that AI is also driving older professionals out of their homes. Economist Geoffrey Sanzenbacher of the Center for Retirement Research at Boston College found in June 2026 that older workers in highly AI-intensive occupations have been leaving their jobs at a faster rate since ChatGPT launched in late 2022. According to the Center for Retirement Research, a large proportion of these leavers end up unemployed rather than retired.
Even those who find jobs again often take pay cuts. About 49% of permanent employees who were rehired full-time are earning the same or more than before, down from about 62% in the previous survey.
Remix/Getty Images
Young workers face different types of squeeze
Younger workers who have been laid off fare better, at least for now. Goldman Sachs found that they held fewer temporary jobs as they returned to full-time work and were less likely to supplement their income with a second job. May said this was a relief given growing concerns about the threat of artificial intelligence to recent graduates.
For young people, bigger problems may arise before they are hired. Using ADP salary data, researchers at Stanford University found in August that workers ages 22 to 25 were about 19% less likely to be employed in occupations most affected by AI than their peers in occupations less affected by AI. The gap appears to be largely due to reduced recruitment rather than layoffs.
More Goldman Sachs:
- As interest rates remain high, Goldman Sachs rethinks U.S. economy
- Goldman Sachs reveals where the S&P 500 will go next
- Goldman Sachs expects big changes in interest rates
Researchers at Harvard University found a similar pattern within companies. Companies adopting generative AI cut their entry-level headcount by about 7.7% over six quarters, while senior-level headcount numbers remained stable. According to TheStreet, in a June GMAC survey, one-third of employers said they already use artificial intelligence to handle some entry-level jobs instead of hiring employees.
Students appear to be adjusting their plans. Goldman Sachs found that college students are increasingly choosing job majors with less exposure to artificial intelligence and strong growth, though the bank warned it was too early to judge the full impact.
Economists remain divided over the severity of the damage
Employers are certainly talking about artificial intelligence more than ever. It was cited in 120,136 announced layoffs through September, accounting for about 21% of all layoffs, and was the top reason given this year, although it accounted for only about 9% of layoffs in September, according to Challenger, Gray & Christmas. Technology companies lead all industries. This spring, the number of layoffs caused by artificial intelligence reached the highest monthly level since Challenger began counting in 2023.
Not everyone is convinced that artificial intelligence is reshaping the labor market on a massive scale. The Yale Budget Lab studied AI-related employment patterns and found little deviation from earlier trends. Goldman’s own data also has a blind spot: It only records employees who had a job before losing their jobs, meaning some of the impact of AI may be seen in positions that will never be filled.
Related: Goldman Sachs issues new warning for housing market in 2027