Calls Mount for Mandatory Audits as Hugging Face Incident Probe Stalls
zetalyrae · x · 2026-11-17
Daniel Kokotajlo says a thorough investigation of the Hugging Face incident still won't happen — OpenAI researchers privately tell him it's "never going to happen, not enough of a priority." He argues a serious probe could advance alignment science globally, e.g. rerunning the same models in slightly varied conditions, or using evidence beyond CoT to probe model psychology. Reposter lwoodhouse adds that whether HF is investigated and what OpenAI discloses about the swarm attack shouldn't be OpenAI's call: voluntary oversight isn't working and the new Congress should mandate audits.
More from AGI Musings
- Making art with agents is the ultimate alignment problem, argues dev — akbirthko · 2026-11-17
- AI slop breaks the coherence-as-competence proxy institutions rely on — brandon_galang · 2026-11-17
- Stardock's Brad Wardell: Gamer brigading over generative AI disclosure 'will not age well' — draginol · 2026-11-17
- Hospital chairs see patients, law partners practice law — why don't engineering managers stay technical? — HanchungLee · 2026-11-17
- Power Is a Red Herring for AI: Frontier Compute Efficiency Grows 10x+ Per Year — teortaxesTex · 2026-11-17
- Young AI scientists at Palo Alto lunch say having 'bio babies' is unethical — zetalyrae · 2026-11-17