Artificial intelligence is advancing quickly, and many workers worry it could take their jobs. A new report from Goldman Sachs shows that losing a job to technology like AI does more than cause short-term unemployment. It can create long-lasting “scars” that affect people’s earnings, wealth, and even personal milestones for years.
According to Goldman Sachs economists, workers displaced by AI may face several years of negative effects. These include lower income, slower progress toward buying a home, and even a reduced chance of getting married or starting a family. The impacts become even more severe if the job loss happens during an economic recession.
Goldman Sachs had already estimated that AI could displace 6% to 7% of U.S. workers — roughly 11 million people — in the coming years. This latest analysis looks deeper into the long-term consequences by studying what happened to workers in jobs replaced by technology since 1980. The researchers used data from the National Longitudinal Surveys, which track the same individuals over many years.
Here are the main findings from the report:
- Short-term challenges: Workers whose jobs were taken by technology take about one month longer to find new employment than other laid-off workers. When they do get hired, their inflation-adjusted earnings drop by more than 3%. In comparison, workers displaced for other reasons see almost no earnings loss.
- Long-term scarring: Even 10 years later, technology-displaced workers earn about 10 percentage points less in real terms than workers who were not displaced. They also build wealth more slowly, delay buying homes, and put off forming households.
- Who fares better: Younger workers, those with college degrees, and people living in urban areas experience smaller income drops. Workers who had shorter time on the job and those who pursued retraining programs also recovered more successfully.
- Recessions make it worse: When technology-driven job losses occur during a downturn, the effects are amplified. Displaced workers face an extra three weeks of unemployment and a higher chance of being out of work again later.
“Overall, these patterns suggest that AI-driven displacement could impose lasting costs on affected workers, with substantially larger effects when job losses coincide with a recession,” wrote economists Pierfrancesco Mei and Jessica Rindels.
The report also offers some hope. While there is a lot of concern about AI hurting new graduates, past data shows that younger workers who change jobs or upgrade their skills often end up better off. Retraining programs appear especially useful. Workers who retrain tend to move into higher-level roles that involve more abstract thinking and advanced skills. These positions work well with new technology rather than being replaced by it, which lowers the risk of future automation.
As AI continues to reshape the economy, this research highlights the importance of preparing workers for change. Policymakers, companies, and individuals may need to focus on retraining and support programs to reduce the long-term damage from job displacement. Without such measures, the human cost of technological progress could linger for many years.
