Andreessen Horowitz partner: data show no mass AI job losses

David George, a general partner at Andreessen Horowitz, wrote an essay arguing that available surveys and working papers show no evidence of economy‑wide job losses from artificial intelligence through early 2026. He cited several recent studies, including an Atlanta Fed survey of roughly 6,000 corporate executives.

The Atlanta Fed survey covered executives in the United States, the United Kingdom, Germany and Australia. More than 90% of respondents reported no AI‑related impact on employment to date. The survey also reported that executives expect larger effects over the next three years, forecasting an average employment reduction of about 0.7% in that period.

National Bureau of Economic Research Working Paper 34984 reached a similar conclusion, reporting that AI adoption had not produced meaningful changes in overall employment to date. The paper noted changes in how work is divided within firms, with routine clerical and administrative tasks appearing more exposed to substitution and AI most often used to assist analytical, technical and managerial work.

Another working paper cited by George found that only about 5% of firms using AI reported any headcount changes. That study also reported signs of capital substitution: 16% of AI‑using companies replaced existing software or equipment with AI‑integrated systems, indicating shifts in investment and technology upgrades.

An April 2026 paper from the Yale Program on Financial Stability’s Budget Lab described economy‑wide AI labor disruption as largely speculative at present. Separate workplace research by Microsoft in 2026 found that workers’ readiness to use AI tools generally outpaced organizational systems and processes, a pattern the company framed as adoption friction rather than immediate displacement.

George rejected forecasts of mass unemployment tied to AI, calling the idea of a widespread job apocalypse a “complete fantasy” in his essay. He also wrote, “Of course AI will absolutely eliminate some tasks and compress some roles,” and argued that productivity changes could alter labor demand.

In his essay George summarized the studies and surveys to show the effects observed through early 2026. The research he cited reports that firms are reallocating tasks, substituting routine work in targeted ways, and changing investment in technology, while broad employment levels have not shown economy‑wide declines.