Workers pushed out by AI are taking longer to land new jobs, and history suggests the damage to their earnings could linger for years
The usual fear about artificial intelligence is a sudden wave of pink slips. Goldman Sachs is pointing to a quieter, slower problem.
Workers who lose their jobs to AI appear to be struggling to find new ones. According to the bank’s analysis, the result may be years of lost earnings, even if the national unemployment rate never budges much.
Goldman Sachs released multiple updates in October 2026 on how AI-driven job losses are reshaping the labor market. The core finding is that AI does not need to cause mass layoffs to leave lasting scars.
The reemployment gap
The headline number is about time. Displaced workers coming from AI-exposed roles took roughly one month longer to secure new jobs than workers from roles without AI exposure.
Age makes the gap worse. Workers aged 35 and older took approximately 5 weeks longer to return to full-time work after displacement. Workers under 35 faced a delay of about 3 weeks.
Goldman’s analysis frames this as a repeat of a familiar story. The pattern tracks historical trends seen during technological shifts over the last 40 years.
Bigger forecasts, bigger stakes
Goldman also revised its long-range displacement forecast upward. The bank now projects that up to 9% of the US workforce, approximately 15 million people, could be displaced over the next decade.
That estimate assumes AI delivers a 15% productivity gain in advanced economies. Goldman’s previous estimate put displacement at 6-7% of the workforce.
Current layoff data suggests the trend is already visible. AI was cited in 21% of all layoff announcements in September 2026, per the data referenced in the research. The tech sector accounted for 29% of total cuts.
The earnings hit may outlast the job search itself. Historical data on displaced workers shows they often see a roughly 3% reduction in real earnings once they are reemployed. Their earnings growth also tends to run about 10 percentage points slower than that of workers who were never displaced.
Why the low-hire, low-fire economy matters
The US labor market has settled into what economists describe as a low-hire, low-fire environment. Overall layoffs can stay modest while the workers who do get cut face a thin market for their skills. If those skills have been made obsolete by AI, the pool of suitable openings shrinks further.
Goldman’s analysis suggests unemployment rates may not see a quick spike under baseline growth scenarios. The pain sits beneath the surface, spread across longer job searches and smaller paychecks.
What this means
For policymakers, the findings shift the question. If the damage shows up mainly in reemployment delays and depressed lifetime earnings, a stable jobless rate could mask real deterioration. Programs aimed at retraining older, mid-career workers look especially relevant given the age gap in the data.
Economists expect displacement effects to intensify during any future economic downturn. Recessions tend to accelerate restructuring, and companies that have been testing AI tools may use a slowdown as cover to cut deeper.
For investors, a 15% productivity gain is a meaningful tailwind for corporate margins. But displaced workers who earn less and take longer to find jobs spend less, which can weigh on consumer demand over time. Sectors with heavy AI exposure, particularly tech, may also face more workforce volatility, as the September layoff data already hints.
The things to watch are the layoff announcement share tied to AI, the reemployment timelines in future labor reports, and whether Goldman revises its 9% forecast again.
