THE ESSENTIALS IN 30 SECONDS

  • Workers likely displaced by AI since 2025 took about one month longer to find a new job and were about 10 percentage points more likely to be unemployed a year later.
  • Workers aged 35 and older needed about five extra weeks to return to full-time work, versus about three weeks for those under 35.
  • AI is not recorded as a cause of job loss in the survey data, so Goldman relies on an estimated AI-displacement score.

A research note from Goldman Sachs economist Pierfrancesco Mei, reported by Fortune on October 30, suggests that American workers who have lost jobs likely affected by artificial intelligence are having a harder time getting back into work than other laid-off workers. Mei describes the effect as similar in size to earlier waves of technological change, which Goldman's past research found to leave lasting marks on careers.

What the note finds

According to Fortune's account, workers likely displaced by AI since 2025 needed about one month longer than other displaced workers to find a new job. They were also roughly 10 percentage points more likely to still be unemployed a year after losing their position.

Age makes a difference. Workers aged 35 and older in the AI-displaced group took about five weeks longer to return to full-time employment, while those under 35 took about three weeks longer. Fortune also reports that, when they are rehired, AI-displaced workers, older ones in particular, often move into roles with similar or higher exposure to AI.

How the estimate was built

The analysis draws on the Current Population Survey's Displaced Worker Supplement, covering more than 15,000 workers who lost jobs between 2016 and 2025. Because the survey never asks whether AI caused a layoff, Goldman built a score combining the AI exposure of the occupation, AI adoption in the industry and whether the position was eliminated. The top 30% on that score were treated as likely AI-displaced. For job losses before 2023, the bank found no significant gap between this group and other displaced workers.

This is an indirect method, and AI Weekly notes the note does not break the delays down by occupation. The results should therefore be read as an estimate rather than a direct count of AI-caused layoffs.

Wider context

Goldman's earlier work, using data back to 1979, found that people displaced by technology saw slower real earnings growth and a persistently higher risk of another spell of unemployment. Fortune places the new findings alongside government data showing 7.4 million workers displaced in 2023–2025, up from 6.3 million in the previous survey. Other research is less conclusive: Fortune cites a Stanford Digital Economy Lab study that links weaker employment among young workers in AI-exposed jobs mainly to reduced hiring, and Yale Budget Lab work that found no clear link between AI exposure and changes in unemployment.