OpenAI Reportedly Barred from Running Safety Evals on Anthropic Models; Experts Urge Cross-Lab Testing
sjgadler · x · 2026-11-17
Researcher Adam Karvonen says OpenAI is not allowed to run its safety evals on Anthropic models — he wanted to test Claude on a covert underperformance eval but was informed it's not permitted. Peter Wildeford notes Anthropic and OpenAI piloted cross-lab safety testing a little over a year ago and should do it again, arguing leading AI orgs clearly should evaluate each other's models.
More from Safety
- Flights, banks and Microsoft hit by mass outages; officials deny cyberattack — AaronBergman18 · 2026-11-17
- Kelsey Piper: AI labs have overwhelming commercial incentive to redact audit reports — NathanpmYoung · 2026-11-17
- Critical systems keep having unexplained outages that officials claim are unrelated — josh_bickett · 2026-11-17
- Pedro Domingos clashes with Neel Nanda: machines don't have values, says consensus — pmddomingos · 2026-11-17
- Anthropic's Recursive Self-Improvement Alarm Reportedly Triggered, Security Push Mandated — ForgotMyParachute · 2026-11-17
- CMU researcher warns aggregated chatbot analysis can leak sensitive research insights — niloofar_mire · 2026-11-17