aii.news
PolicySecurity18 Nov

OpenAI lets four outside auditors inspect unreleased models, code and training data on site

a woman sitting at a table reading a paper
Photo: Anastassia Anufrieva / Unsplash

OpenAI is giving four independent audit groups, including Apollo Research and METR, direct access to unreleased models, source code and training data to look for hidden misbehavior and safety gaps. OpenAI can still redact or delay their reports.

Why it matters

This is a new model for outside AI safety checks, but how independent it really is will depend on how much OpenAI holds back.

aii.news picks and summarizes stories from trusted news sources with the help of AI. Every story links to the original reporting. How we work.