Goldman Lifts AI Displacement Forecast to 9% of US Workers

Goldman Sachs now estimates that about 9% of US workers, roughly 15 million people, could be temporarily displaced by AI over the coming decade, an upward revision from a prior 6-7% estimate. The revision, from Goldman economist Joseph Briggs, assumes AI eventually lifts US productivity by 15% in developed economies, with an additional 3% of workers facing slower hiring.

The timing matters at least as much as the headline figure. 'AI-related labor disruptions could well amplify the broader business cycle,' Briggs writes, noting that 'automation-related labor disruption is limited when the overall economy is healthy but accelerates during cyclical slowdowns.' Goldman's look back at routine occupations, clerks, secretaries and factory operators, found employment steady through expansions and sharply down in the recessions of the early 1980s, 2001, 2008 and 2020.

AI is already removing between 10,000 and 15,000 jobs a month from US payroll growth, concentrated in tech, management consulting and graphic design. Only 23% of US establishments have integrated AI into regular workflows so far, and AI-attributed layoffs peaked in spring 2026 before cooling significantly. Goldman's analysis of 25 occupations hit hardest by prior technology waves showed average job losses of 25% after five years once the tech made the work redundant.

The macro read lands in a stretch our Jobs tracker has been clocking in real time, with 102 AI-adjacent labor stories logged in the last 90 days, including eBay's second 2026 layoff round cutting 9% of its workforce.