Anthropic gives METR "employee-like" access to assess AI risks
Anthropic has granted nonprofit evaluator METR "employee-like access" to its models, hardware and internal systems, embedding six staff in its office. METR will publish its findings without Anthropic's veto, though the company can require redaction of customer data and model architecture details.
- Six METR staff received company hardware, office space and internal tool access
- Anthropic does not pay METR and cannot veto publication of its analysis
- Anthropic may require redaction of customer data and model architecture details
- The move comes as Anthropic preps an IPO at a valuation above $2 trillion
Read next
AI