Anthropic says Claude now leads a quarter of work building its next AI models

Anthropic ​said that Claude is not operating fully autonomously in any part of ​the work measured, but its contribution to the technology's development has ⁠risen rapidly from 1% in March on a scale developed by Epoch AI, ​an independent nonprofit that tracks the technology.

The figures offer one of the most detailed ​public windows into how leading companies in the industry are developing the technology amid growing concern that AI could soon reach a point at which it could feed its own development ​with little input from humans.

Researchers have warned that as AI agents become more ​autonomous, they may develop behaviors that diverge from their creators' intentions and become harder to monitor ‌or ⁠control. On Wednesday, the company's rival, OpenAI, said it would begin regularly publishing reports on unexpected or unauthorized AI behavior, while disclosing six reports of unexpected or concerning model behavior.

Anthropic said about 30,000 AI agents were doing research and engineering work on ​its main internal platform ​at any one ⁠time in August. It added that every action the agents take is screened before it is carried out, and of more ​than a billion decisions that month, about one in 47,000 ​was blocked.

The ⁠company also disclosed that about 6% of the computing power Anthropic used for AI research went to safety work in a sample week in July, with the figure ⁠rising ​to 12% for research carried out by AI itself. ​It said the figures were conservative as the computing power that advanced safety and capability equally was ​counted as capability work.