Goldman Sachs Expects AI May Temporarily Replace About 9% of U.S. Labor Over the Next Decade
Goldman Sachs economists Joseph Briggs and their research team said in updated forecasts around October that, under the base case scenario (with productivity in developed economies eventually rising by about 15%), artificial intelligence could temporarily replace about 9% of U.S. labor over the next decade, equivalent to about 15 million jobs. This proportion is higher than the previous estimate of about 6% to 7%.
The research is based on historical data: when technology-driven productivity rises by 1 percentage point, the job displacement rate over the next two years typically increases by about 0.5 to 0.6 percentage points. Goldman Sachs said that industries such as technology, management consulting and graphic design have already seen AI-related job losses of about 10,000 to 15,000 per month, while only about 23% of U.S. institutions currently integrate AI into regular workflows. The company also expects AI to create new jobs in the long run to partially offset the losses, making the labor market shock temporary; if job losses are spread across ten years and most people are reemployed within one year, the peak impact on the unemployment rate would be limited.
The forecast emphasizes that the impact is related to the economic cycle: automation-related labor shocks are relatively limited when the economy is healthy, but may accelerate during recessions. Goldman Sachs chief economist Jan Hatzius and others also said that under a base path close to potential growth, AI-related labor disruptions are more likely to emerge gradually rather than as a sudden shock. Source: public information.
Original sources:
Title: AI Layoffs Could Displace Nearly 10% of Workers in Coming Years, Goldman Says
Publisher: Business Insider
Publication date: 2026-10-23
URL: https://www.businessinsider.com/ai-layoffs-job-market-prediction-goldman-sachs-2026-10
Source type: media report (citing Goldman Sachs research)
Verification status: read
Supporting content: under the base case, up to about 9% of labor replaced, about 15 million people, revised up from 6%-7%Title: Goldman Sachs says AI will eliminate nearly 15 million jobs
Publisher: Crypto Briefing, etc. (aggregating Goldman Sachs research)
Publication date: 2026-10-30
URL: https://cryptobriefing.com/goldman-sachs-ai-eliminate-15-million-jobs/
Source type: secondary report
Verification status: cross-checked across multiple sources
Supporting content: 9% or about 15 million jobs, productivity assumption, current monthly job-loss data in industries
ABAB AI Insight
The core of Goldman Sachs’ update is that its estimate of AI’s “temporary displacement” of the U.S. labor market has been raised from the previous 6%-7% to as much as about 9% (about 15 million people), based on the empirical relationship between productivity gains and historical job displacement rates. This figure describes the scale of jobs that may leave their current positions over the next decade, not a permanent net loss of 15 million jobs. The research also says AI will create new jobs in the long run, and if job losses are spread out and reemployment is fast, the peak impact on the unemployment rate will be limited. At present, technology, consulting and design industries have already seen about 10,000 to 15,000 related jobs lost per month, but the overall AI integration rate is still only about 23%.
The forecast explicitly ties the shock to the business cycle: when the economy is healthy, automation-related layoffs are more restrained, while during recessions they may accelerate. This is consistent with the historical pattern in which job adjustments from technological change often concentrate in downturns. For markets and policy, the key is not the single headline number of 15 million, but that the displacement is temporary, the speed of reemployment, the pace of new job creation, and whether the impact is concentrated in specific occupations and age groups. Goldman Sachs itself said the shock is more likely to emerge gradually rather than collapse suddenly.
This estimate is a model-driven scenario analysis, relying on the assumed magnitude of productivity gains, historical elasticity coefficients and reemployment assumptions. The actual path will be affected jointly by the speed of corporate adoption, regulation, retraining and the macroeconomic environment. Public information shows that Goldman Sachs has raised the figure and provided methodological support, but it has not said that 15 million jobs will be permanently eliminated.
ABAB News · Cognitive Law
9% and about 15 million refer to the scale of “temporary displacement,” not a promise of a permanent net reduction in employment.
Historically, automation-related job adjustments more often accelerate during recessions rather than being evenly distributed.
Current AI-related monthly job losses have already appeared, but the overall institutional integration rate remains low, and the shock is still at an early stage.