Andreessen Horowitz Partner: Data Shows Little AI Job Loss

David George, general partner at Andreessen Horowitz, wrote in an essay that surveys and working papers show little evidence of economy-wide job losses from artificial intelligence through early 2026 and called the idea of mass AI-driven unemployment “a complete fantasy.”

He cited an Atlanta Fed survey of roughly 6,000 corporate executives in the United States, the United Kingdom, Germany and Australia. More than 90% of managers reported no AI-related impact on employment so far. The survey found executives expect AI to reduce employment by about 0.7% over the next three years.

George referenced National Bureau of Economic Research Working Paper 34984, which found AI adoption has not led to meaningful changes in overall employment to date. The paper reported firms are reorganizing tasks internally; routine clerical and administrative duties are most exposed to substitution, while AI is more often used to support analytical, technical and managerial work.

A separate working paper examined firms using AI and reported about 5% had made headcount changes. The same paper found 16% of AI-using firms had replaced software or equipment with AI-integrated solutions, a pattern the authors described as capital substitution rather than widespread layoffs.

The Yale Budget Lab’s April 2026 paper stated that economy-wide labor disruption from AI remains largely speculative. Microsoft’s 2026 workplace study found worker readiness for AI tools currently outpaces many organizations’ capacity to deploy them, a gap researchers linked to adoption friction.

George wrote that while AI will eliminate some tasks and compress some roles, claims of permanent, economy-wide unemployment are misleading. He wrote, “Of course AI will absolutely eliminate some tasks and compress some roles,” and described claims of permanent unemployment as “unhelpful marketing, bad economics and worse history.”

The studies George cited report limited employment effects through early 2026, changes in task allocation within firms and increased investment in AI-enabled systems. None of the papers documented large-scale, permanent job losses across the economies they examined.