Claude now leads 26% of Anthropic AI research: Report

The Washington Post reported Thursday that an internal review found Claude "leads" approximately 26% of Anthropic's AI research and development work, rising from effectively zero as recently as February. The designation indicates the model can complete most tasks from high-level prompts while humans provide supervision and direction, according to the report.

Anthropic stated that Claude is not currently operating fully autonomously in any measured R&D area. However, the company noted that more than 90% of its research now involves Claude either collaborating with researchers or taking the lead role.

Path toward recursive improvement

The company said the sharp increase could signal progress toward "recursive self-improvement," a scenario in which AI systems become capable of developing more advanced successors with increasingly limited human involvement. Anthropic did not specify how close current systems are to reaching that threshold.

The artificial intelligence developer disclosed that approximately 30,000 AI agents were conducting research and engineering work on its most-used internal platform as of August. The company added that it maintains monitoring systems designed to detect and flag potentially problematic agent activity for human review.

Industry transparency calls

Anthropic has urged other leading AI developers to publish similar measurements, arguing that greater disclosure would allow governments, researchers and the public to better track the pace of AI development. The appeal comes after CEO Dario Amodei called for coordinated slowdowns in frontier AI development to provide additional time for safety measures.