Goldman Sachs: AI's One-in-11 Job Shock Is Waiting on a Recession

Goldman Sachs economist Joseph Briggs ran the numbers and came back with a headline-grabber: up to 9% of workers could be displaced in the AI transition, with slower hiring dragging employment down another 3%. That's roughly one in 11. But the sharper finding in his note published Thursday is about timing — Briggs argues the layoffs aren't on AI's clock, they're on the economy's. "Automation-related labor disruption is limited when the overall economy is healthy but accelerates during cyclical slowdowns," he wrote. Translation: don't expect the bloodletting until the next downturn gives companies an excuse to restructure. Briggs calls it a feedback loop — "AI-related labor disruptions could well amplify the broader business cycle." It's the second time Goldman has made this call in a year, from two different teams. Back in October 2025, U.S. economists David Mericle and Pierfrancesco Mei predicted "jobless growth" as the likely new normal and warned the full AI labor impact "might not become apparent until a recession hits." The evidence is decades deep. Briggs tracked the share of Americans in routine jobs — clerks, secretaries, machine operators, mechanics — and found it holds steady during expansions, then drops in hard steps during downturns. The line fell through the early 1980s, 2001, 2008 and 2020 recessions and never climbed back. Across 25 occupations hit by past technology, employment fell an extra 4% over five years for every point the national unemployment rate climbed. The math behind the 9% figure: every 1% gain in technology-driven productivity has historically pushed unemployment up 0.3 to 0.4 points at its peak, fading to zero after two years, while lifting the cumulative gross job-loss rate by 0.5 to 0.6 points. Multiply 0.6 by Goldman's projected 15% AI productivity boost and you land at 9. Briggs also looked at 25 hard-hit jobs — photo processors, movie projectionists, proofreaders, file clerks, bank tellers — where employment fell 25% on average in the five years after decline began. He calls those "likely upper bounds." Briggs's own estimate has drifted. In August 2025, he and economist Sarah Dong put displacement at 6% to 7%, inside a 3% to 14% range. By June, on Goldman's Exchanges podcast, he was saying 9% — about 15 million workers over a decade — and describing them as "reallocated to new positions." Thursday's note calls them displaced. He still frames the pain as temporary, saying the annual unemployment increase would likely stay under one percentage point. For now, the alarm bells are quiet. AI-attributed layoffs in Challenger's tracking have pulled back sharply from their spring peak, and news coverage of AI job losses has dropped to early-2026 levels. Briggs notes job growth in the most AI-exposed industries is running about 10,000 below trend over three months, roughly offset by AI-construction hiring. Even Anthropic CEO Dario Amodei has softened his tone; after spending 2025 warning AI could wipe out half of entry-level white-collar jobs, he reached for the Jevons paradox onstage with JPMorgan Chase CEO Jamie Dimon in May, per Fortune. His one caveat is Briggs's whole point: "AI is moving faster than all these previous technologies."

entities: [