AI could displace one in 11 workers, Goldman says, but the layoffs are likely waiting for a recession

Goldman economist Joseph Briggs estimates AI could displace up to 9% of workers under a forecast that AI lifts productivity by 15%, with slower hiring potentially weighing on employment by a further 3%. Goldman’s historical analysis finds technology-related job disruption intensifies during downturns; for 25 occupations affected by past technologies, employment fell an additional 4% over five years for each percentage-point rise in national unemployment. Near-term AI-related labor headwinds remain modest, with only 23% of U.S. establishments using AI in regular workflows, Briggs said.

Analysis

The investable signal is a recession amplifier, not an immediate AI-layoff shock. In a downturn, weak demand gives management cover to cut roles and accelerate automation; resulting wage income loss can further weaken consumer spending, while productivity gains support margins for firms able to retain demand. Headline unemployment may understate damage if displaced workers exit the labor force, so monitor payroll separations, hours worked, and labor-force participation alongside the jobless rate.

The 6–18 month risk is asymmetric for labor-intensive businesses and consumer-facing firms: weaker hiring and household income can pressure volumes before automation savings show up in reported margins. Conversely, automation vendors may benefit structurally, but their near-term revenue still depends on corporate budgets and adoption—not announced intent. A recession could therefore hurt both labor-intensive customers and AI infrastructure spending before any productivity payoff.

This is not a GS-specific earnings catalyst; treat the research as a macro risk input, not evidence of Goldman Sachs’ own exposure. The contrarian point is that subdued layoffs do not disprove displacement, but the historical analogues are selected episodes and do not establish the scale or speed of AI’s net employment effect. Without valuation, exposure, and adoption data, avoid chasing an AI-loser basket today.

Key Decisions for Investors

• No immediate position in GS on this item alone. Reassess if labor-market deterioration begins to translate into weaker consumer credit or materially revised bank earnings guidance.
• Set a conditional relative-value alert: on a clear rise in unemployment claims and declining hours worked, consider short labor-intensive consumer discretionary exposure (e.g., XLY) against a diversified automation-enabler basket. Size modestly; recessionary capex cuts can also hit AI suppliers.
• Track separations, job openings, hours, participation, and AI-related spending/adoption over the next 1–3 months. The thesis strengthens if layoffs rise while productivity remains firm; it weakens if hiring stabilizes and productivity gains fail to broaden.
• Falsification: sustained labor-market resilience alongside stable consumer spending, or evidence that AI adoption is augmenting roles without a corresponding rise in separations, would argue against treating this as a near-term earnings or recession trade.