According to Goldman's analysis, workers aged 35 and above who were likely displaced by AI needed roughly five additional weeks to return to full-time employment compared with other laid-off workers. Those younger than 35 experienced a smaller delay of approximately three weeks.
Younger workers also tended to move through fewer temporary positions before securing permanent employment and were less likely to need an additional job alongside their new full-time role. Mei suggested that their relative flexibility could help them adjust to changing workplace demands.
The analysis draws on more than 15,000 people who involuntarily lost employment between 2016 and 2025, using the federal Current Population Survey's Displaced Worker Supplement.
Because government employment surveys do not directly identify layoffs caused by AI, Goldman developed a classification method combining occupational exposure to the technology, industry adoption rates, and whether a worker's position had been eliminated. The researchers categorized the highest-scoring 30% as likely AI-displaced.
That distinction is important: the study identifies workers whose layoffs were likely connected to AI rather than establishing that the technology directly caused every job loss.
Goldman found no statistically significant difference in reemployment outcomes between the two groups before 2023, when AI adoption began accelerating. The more recent divergence suggests that workers in highly exposed positions are encountering additional difficulties finding replacement jobs.
The bank compared those results with earlier research examining technology-related layoffs dating back to 1979. That historical analysis found that displaced workers often experienced slower inflation-adjusted earnings growth and remained more vulnerable to subsequent unemployment.
The long-term consequences varied considerably. Younger workers, college graduates, and people living in urban areas generally suffered cumulative earnings losses roughly half those experienced by other technology-displaced employees over the following decade.
Workers with shorter tenure at their previous employers also recovered more successfully, potentially because their skills were easier to transfer between occupations.
Goldman's latest findings indicate that workers displaced from AI-exposed positions are more likely to change both industries and occupations when seeking new employment. However, many eventually enter jobs with comparable or greater exposure to AI, particularly among older workers.
Broader federal employment statistics provide additional evidence of the challenges facing displaced employees, although those figures do not isolate AI-related layoffs.
The Bureau of Labor Statistics recorded 7.4 million displaced workers between 2023 and 2025, compared with 6.3 million in the preceding survey period.
Among employees who had spent at least three years in their previous jobs, 72.9% of those aged 25 to 54 had returned to employment by January 2026. The corresponding rates were 57.3% for workers aged 55 to 64 and 38.6% for those 65 and older.
Recovering previous earnings has also become more difficult. Approximately 49% of long-tenured workers who found another full-time position earned at least as much as they had before losing their jobs, compared with about 62% in the earlier survey.
Meanwhile, employers are increasingly identifying AI as a factor behind workforce reductions. Outplacement firm Challenger, Gray & Christmas reported 120,136 announced job cuts attributed to AI through September, representing approximately 21% of all announced reductions and making AI the most frequently cited reason.
Other research, however, presents a more complicated picture of AI's influence on employment.
An August study from Stanford's Digital Economy Lab found that employment among workers aged 22 to 25 in highly AI-exposed occupations was approximately 19% below the level implied by employment trends among similarly aged workers in less-exposed positions.
Researchers Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen attributed the difference primarily to reduced hiring of younger employees rather than increased layoffs. They found no comparable employment gap among experienced workers and did not identify evidence of widespread AI-driven displacement across the broader economy.
Separately, the Yale Budget Lab found no clear relationship between changes in employment or unemployment and measures of occupational AI exposure, automation, or augmentation.
These findings examine different aspects of the labor market. Stanford's research focuses partly on hiring opportunities for younger workers, while Goldman examines what happens after people lose existing jobs. Goldman's dataset does not capture new entrants who struggle to secure their first positions.
The bank also cautioned that its evidence remains preliminary, particularly when assessing differences by education and geography. Its latest analysis does not provide a breakdown of the additional unemployment duration by occupation.
For now, Goldman's research points to a measurable disadvantage among workers classified as likely displaced by AI, especially those aged 35 and older. Whether those early employment difficulties develop into the persistent earnings losses documented during earlier technological shifts remains an open question.
This analysis is based on reporting from Fortune.
Image courtesy of Unsplash.
This article was generated with AI assistance and reviewed for accuracy and quality.
