Anthropic Grants METR Access for AI Safety
Anthropic has granted nonprofit METR employee-level access to its internal systems to evaluate artificial intelligence safety. The company is also negotiating similar partnerships with Hugging Face and other firms.
Xurve View
Insights:
Xurve View
Anthropic granted nonprofit METR employee-level access to its systems on Thursday as AI labs turn to external evaluators to build safety guardrails without delaying product rollouts. The developer of Claude models is expanding its third-party review framework after previous discussions stalled over trade secret protections. The outcome of these negotiations will shape how external researchers scrutinize advanced artificial intelligence models.
