Key facts
- Anthropic's AI Claude leads 26% of the company's AI research and development work.
- AI collaborated with humans on over 90% of Anthropic's research work as of August.
- Claude's contribution to AI development rose from 1% in March to 26% in August.
- About 30,000 AI agents were engaged in research and engineering on Anthropic's platform in August.
- Approximately 6% of Anthropic's computing power for AI research was dedicated to safety work in July.
- About 12% of computing power for AI research conducted by AI itself was dedicated to safety work in July.
Anthropic, the artificial intelligence company, announced that its AI assistant, Claude, is now leading 26% of the company's internal research and development efforts. This figure represents a significant increase from just 1% in March, according to data published by the company on Thursday.
Claude is not operating fully autonomously in any measured part of the work, but its contribution has risen rapidly. The company also stated that AI collaborated with humans on more than 90% of its research work as of August. These figures offer a detailed public view into how leading AI companies are developing the technology, amid growing concerns about AI autonomy and potential divergence from human intentions.
In August, approximately 30,000 AI agents were engaged in research and engineering work on Anthropic's main internal platform. The company noted that every action taken by these agents is screened before execution, with about one in 47,000 decisions being blocked out of more than a billion decisions made in August.
Furthermore, Anthropic disclosed that about 6% of its computing power used for AI research was allocated to safety work in a sample week in July. This figure rose to 12% for research conducted by AI itself. The company stated these figures were conservative, as computing power used equally for safety and capability advancements was counted as capability work.