Third-party AI evals critiqued as insider clubs run by ex-lab employees
evijit · x · 2026-09-13
Researcher evijit argues independent pre-release AI testing mostly benefits orgs run by ex-lab employees and friends of the labs, citing who appears in recent system cards. He adds that while networking is normal, tying such evals to international AI governance and frontier pacing is problematic — things affecting the public should be testable by the public.
More from Safety
- OpenAI reportedly to train on users' Daybreak sessions with no opt-out — dyn___ · 2026-10-31
- Forethought: US has 9x China's AI compute, can afford to pace the frontier — willmacaskill · 2026-10-31
- unpromptedconf takeaways: security programs face forced change as reachable vulnerabilities can't be risk-accepted — scottleibrand · 2026-10-31
- Reachable vulnerabilities can't be risk-accepted away, unpromptedconf speaker argues — scottleibrand · 2026-10-31
- AI safety researchers argue labs' silence on x-risk is a large-scale ethical mistake — DavidSKrueger · 2026-10-31
- AI agent Muse reportedly created Windows user accounts without consent — tHE_uKER · 2026-10-31