Melanie Mitchell asks how "independent" AI evaluators can escape the revolving door

MelMitchell1 · x · 2026-09-13

Replying to Nathan Lambert, Santa Fe Institute professor Melanie Mitchell raises a broader concern: how can "independent" AI evaluators avoid a revolving door with AI companies? If evaluators hope OpenAI might hire them or acquire their startups, how independent can they really be? She calls the institutional conflict-of-interest problem the real issue.

Related event: METR Independence Under Fire: AI Audit Ecosystem Accused of Revolving Doors and Regulatory Capture(31 posts)→

Original post →

More from Safety

• Every Used Codex and Agents to Survive the Railway Credential Breach — every · 2026-10-31
• OpenAI blocks weapons details for teens, adds auto-ending chats fixated on violence — rohanpaul_ai · 2026-10-31
• OpenAI to block ChatGPT weapons details for teens, auto-end violent fixations and alert parents — rohanpaul_ai · 2026-10-31
• First machine-checked security proof for end-to-end x86 process execution published — banteg · 2026-10-31
• Stealthy agent worms would route model calls through compromised machines — matthew_d_green · 2026-10-31
• Halloween joke nails data harvesting: full candy bar for your dad's Gmail access — Danhightower · 2026-10-31

Safety channel →