Anthropic’s Economist Says AI Has Not Yet Shocked the Labor Market

Anthropic’s chief economist Peter McCrory argues, based on internal data, that artificial intelligence has not yet delivered a major blow to the U.S. labor market.

Findings From In‑House Research

McCrory’s synthesis of 18 months of internal analysis yields the following takeaways:

■Unemployment rate in June stood at 4.2%, the level the Federal Reserve deems full employment.

■Job openings roughly match the number of unemployed, while prime‑age employment hits multi‑decade highs.

■Workers in AI‑heavy roles show no statistically significant rise in unemployment compared with low‑exposure occupations.

Dario Amodei’s Shifting Warning Signals

Amodei’s public statements have oscillated over the past two years:

■May 2025: Forecasted 10‑20% unemployment, claiming half of entry‑level white‑collar jobs could vanish.

■January 2026: Described AI as a “general labor substitute,” urging policies such as universal basic income.

■May 2026: Invoked the Jevons paradox, arguing that automating 90% of a job expands the remaining 10% into 100% of what people do, boosting productivity tenfold.

■June 2026: Re‑escalated, suggesting that significant, enduring job loss may be an “intrinsic property” of the technology.

Real‑Time Labor Market Indicators

Data from the Bureau of Labor Statistics (BLS) corroborates McCrory’s narrative:

■AI‑exposed occupations (e.g., technical writers, data entry, customer support) show a softening in hiring for younger workers.

■Stanford’s “canaries in the coal mine” study flags these groups as early‑risk cohorts.

■BLS projects slower growth through 2034 for the aforementioned occupations.

Technology, the Jevons Paradox, and Conflicting Scenarios

McCrory highlights AI’s “jagged” capability profile as a barrier to full substitution:

■No O*NET taxonomy job is entirely covered by Claude’s task set.

■Users treat Claude as a “thought partner,” and domain expertise improves error‑catching when the AI falters.

■The Jevons paradox requires market and workforce adjustment time, whereas Amodei repeatedly claims AI is moving faster than any prior general‑purpose technology, undermining that equilibrium.

Expert Comment (Bora Yalın): Short‑term spikes in unemployment from AI appear unlikely. The softening of entry‑level hiring, however, signals the early stage of a skill‑biased reallocation. Investors should assess productivity gains and margin risk in AI‑heavy firms separately, while policymakers may need to phase in wage insurance and basic income as technology adoption unfolds. Markets will feel AI’s “bigger pie” effect, but the winner‑loser dynamics for workers remain uncertain.