AI Displacement Could Affect 6–7% of US Workforce, Lower Earnings
Goldman Sachs economists estimate AI could displace 6–7% of US workers (about 11 million), who face one-month longer job searches and over 3% bigger earnings declines short term. A decade after displacement, affected workers earn 10 percentage points less in real terms, delay homeownership and face 5% more joblessness during recessions.
1. Worker Displacement Estimates
Economists project that AI could displace 6–7% of the US workforce, equal to approximately 11 million workers, by automating roles across multiple sectors.
2. Short-Term Labor Outcomes
Displaced workers take on average one month longer to find new jobs and suffer more than a 3% drop in inflation-adjusted earnings compared with non-displaced peers.
3. Long-Term Scarring Effects
Ten years after losing jobs to technology, affected workers’ real earnings remain about 10 percentage points below non-displaced workers, with slower wealth accumulation, delayed homeownership and household formation.
4. Recession Intensifies Impact
When displacement coincides with a recession, unemployment duration increases by three additional weeks and the likelihood of subsequent joblessness rises by about 5%, though younger and retrained workers fare relatively better.