Policy
OpenAI opens its AI systems to four independent on-site auditors for the first time
The Decoder · By Manuel Uth ·
The Brief
- OpenAI is giving four independent audit organizations, including Apollo Research and METR, direct access to unreleased models, source code, and training data.
- The stated goal of the embedded on-site reviews is to uncover hidden misbehavior and safety gaps.
- OpenAI can redact or withhold some information from the auditors, according to the source.
Why it matters
It marks a first-time step toward external, on-site auditing of frontier AI systems, which may shape how safety evaluations of AI models are done and reported.
Who’s affected
- AI safety researchers
- Developers building on OpenAI models
- #ai safety
- #openai
- #model auditing
- #third-party evaluation
- #ai governance
OpenAI is giving four independent audit organizations, including Apollo Research and METR, direct access to unreleased models, source code, and training data for the first time. The embedded reviews are designed to uncover hidden misbehavior and safety gaps. OpenAI can redact or…