Goldman Sachs sees 15 million US workers displaced by AI

Goldman Sachs now estimates more than 9% of the US workforce — roughly 15 million workers — could face job displacement over the next decade as AI adoption accelerates.

Goldman Sachs raised its estimate of AI-driven US job displacement to more than 9% of the workforce, or about 15 million workers, as productivity gains from the technology begin to show up in employment data.

"Despite our expectation that AI-related job losses will lead to a meaningful amount of labor displacement, we continue to expect that labor market headwinds will be temporary," Joseph Briggs, senior global economist at Goldman Sachs, said in the report published June 25.

The revised forecast marks an increase from the firm's earlier estimate of 6% to 7% of workers. The new methodology measures the flow of workers leaving existing jobs rather than the stock of unemployed workers. Every 1% increase in technology-driven productivity has historically raised the job destruction rate by 0.5 to 0.6 percentage points over the following two years, Briggs found.

The question for investors and policymakers is whether the US economy can create new jobs fast enough to absorb displaced workers. The economy typically generates 25 million to 35 million new positions annually, but early signs of strain are already visible: employment in AI-penetrated sectors is shrinking by about 11,000 jobs per month, and recent college graduates face an unemployment rate higher than other education groups.

Coding agents already cheaper than humans

Goldman's analysis shows that programming AI agents have already reached cost parity below human developers, driven by declining token prices. As inference costs continue to fall, the range of roles where automation is economically viable will expand, the bank said. Companies citing AI as a reason for layoffs have increased sharply in recent months, according to Challenger, Gray & Christmas data. Andy Challenger, labor and workplace expert at the outplacement firm, said AI is now "the leading reason companies give for cutting jobs."

New jobs vs. old jobs — the race that defines the decade

Goldman expects AI to generate new employment over time through emerging occupations and increased specialization. Box Inc. Chief Executive Officer Aaron Levie said AI has created 13 new job categories at his company, including AI architects and model evaluators.

But the pace of creation may not match the pace of destruction. Daron Acemoglu, an Institute Professor at the Massachusetts Institute of Technology and a Nobel laureate in economics, said "no general law of economics guarantees that the pace of job creation will match the pace of job destruction." He expects AI to have a modest net negative impact on employment over the next five years.

Neil Thompson, director of the FutureTech research project at MIT's Computer Science and Artificial Intelligence Laboratory, described AI's impact as "a rising tide" rather than "a crashing wave," suggesting businesses and workers will have time to adapt. Goldman's Elsie Peng said AI augmentation has created jobs, but not enough to fully offset losses from substitution, resulting in a small net drag on the labor market.

Assuming AI adoption unfolds over a 10-year period and most displaced workers find new employment within a year, Briggs estimates the peak impact on the US unemployment rate would be less than one percentage point — significant but manageable within the context of the economy's long history of technological change.

This article is for informational purposes only and does not constitute investment advice.