Anthropic said that from May 2026 its agents logged more work hours than human employees. It said employees spend about 22% of their time managing Claude agents, more than twice the share a year earlier. The company also reported that entry-level technical hires fell from nearly 30% to 9% since March 2025, while the junior share of headcount fell from about 21% to 14%.
Morgan Frank of Case Western Reserve University told the Guardian that AI-exposed US job markets had weakened from early 2022, before ChatGPT or Claude, and said post-pandemic cuts or “AI-washing” could also explain losses. Yale’s Julian Posada said the report lacked independent outside auditing. Distributed AI Research Institute engineer Dylan Baker questioned its description of humans as a bottleneck.
The source does not establish that Claude caused Anthropic’s hiring changes, that the company’s internal experience generalizes to other employers, or that AI caused wider labor-market deterioration. Anthropic did not respond to the Guardian’s request for comment.