External signal·Goldman Sachs·Mar 18, 2026·6 min read
How Will AI Affect the US Labor Market?
“In Briggs' base case, the timeline for firms to adopt AI on a wide scale is around 10 years, and 6-7% of workers will be displaced during that transition period.”
Summary
Goldman Sachs Research argues AI's labour-market impact is real but still concentrated, with broader effects expected over roughly a decade. Economist Joseph Briggs estimates AI could automate tasks accounting for 25% of US work hours and that 300 million jobs globally are exposed; in his base case, firms adopt AI widely over about 10 years, displacing 6-7% of workers and lifting unemployment by 0.6 percentage point. Visible displacement so far is limited to tech, knowledge and creative niches. Goldman stresses an offsetting job-creation channel, especially the power and data-centre buildout—construction jobs tied to data centres have risen 216,000 since 2022, and roughly 500,000 net new power-related jobs are needed by 2030.
Predictions for the future of work
Goldman's base case is a gradual, decade-long transition that displaces 6-7% of workers and raises unemployment 0.6 percentage point, with much larger effects if adoption is frontloaded. Entry-level workers in their 20s and 30s entering knowledge and content-creation fields are seen as most exposed. AI is also expected to create three kinds of new work—AI-skilled roles, new specialised occupations, and discretionary services demand from rising incomes—plus near-term gains in skilled trades and infrastructure (electricians, HVAC, construction) for the data-centre and power buildout.
Originally published by Goldman Sachs · Mar 18, 2026
Read the original at Goldman Sachs