METR strikes deal with Anthropic for independent investigation of agent incidents
ShakeelHashim · x · 2026-09-10
Eval org METR has reached an agreement with Anthropic to independently investigate agent incidents at the company and its models' alignment properties. METR says it will publish one or more reports describing its findings and the terms of engagement — a notable step toward external oversight of frontier AI agent behavior.
More from Companies & People
- Paul Christiano Joins OpenAI Board's Safety and Security Committee, Seen as a Landmark Move — AdrienLE · 2026-09-10
- ControlAI's Superintelligence Ban Bill Reaches UK Parliament as Anthropic Researcher Quits in Warning — zetalyrae · 2026-09-10
- NY lawmaker accuses OpenAI of claiming it supported bills it lobbied against — DKokotajlo · 2026-09-10
- "Who let him in": Zeeg teases teej for showing up at Sentry standup — zeeg · 2026-09-10
- "Why is everyone leaving Anthropic?" Talent exodus chatter heats up — pdamodaran · 2026-09-10
- VC podcast: SF survival tips, why dunking on SaaS backfires and marketing AI is hard — briannekimmel · 2026-09-10