White House Accord On Super Intelligence

Anthropic Gives METR Employee-Like Access for Independent AI Safety Reviews

Anthropic gives six METR staff employee-like access to assess frontier AI risks, publish findings and challenge redactions under a new oversight model.