Anthropic Reveals Internal Metrics on AI Self-Development

Anthropic has published internal measurements tracking how much AI contributes to building future AI systems, human oversight of AI agents, and compute resources powering model development. As of August 2026, Claude leads 26% of Anthropic's AI R&D work, with over 90% of work involving AI collaboration or higher, though no tasks are fully autonomous. The company plans to embed independent third-party evaluators with access to internal systems to verify safety practices and monitor key metrics. Anthropic frames these disclosures as a starting point for external monitoring of AI development pace, and calls for other frontier labs to adopt similar transparent reporting using common methodologies.