Independent AI Audits

Anthropic Gives Six METR Researchers Employee Access to Audit AI Safety

Anthropic gives six METR researchers employee-level access to independently assess its AI models, safety controls, incidents and risk management from inside.