Anthropic Economist Disputes CEO's AI Job Loss Warnings

Anthropic's head of economics, Peter McCrory, published an essay this week asserting that artificial intelligence has not caused a material increase in US unemployment. This data-driven perspective directly challenges the more alarmist predictions made by Anthropic CEO Dario Amodei regarding an imminent "white-collar bloodbath."

Amodei has previously articulated dire forecasts about AI's labor market impact. In May 2025, he predicted AI could eliminate half of all entry-level white-collar jobs and drive unemployment to 10%-20% within one to five years, urging stakeholders to acknowledge the risks. He reiterated these concerns in a January 2026 essay, "The Adolescence of Technology," describing AI as a "general labor substitute" that would displace work across skill levels and potentially create a permanent underclass. His stance has evolved, with a May statement suggesting automation might act as an output multiplier, referencing the Jevons paradox. However, by June, Amodei re-escalated his concerns, labeling significant job loss as an "intrinsic property of the technology" and advocating for government interventions like wage insurance and universal basic income.

McCrory's analysis, which synthesizes 18 months of Anthropic's internal economic research, presents a contrasting view. He stated, "We don’t see significant impact of AI on the U.S. labor market," at least not to date. The US unemployment rate was 4.2% in June, a figure the Federal Reserve considers indicative of full employment. Concurrently, job openings were approximately equal to the number of unemployed individuals, suggesting a balanced labor market.