theaicatalog
OpenAI ยท Policy & safety

OpenAI signs four independent assessors for embedded safety reviews

OpenAI has signed agreements with Apollo Research, METR, Transluce, and Goodfire for the first phase of its embedded assessments. These assessments give outside experts deep access to the company's internal models and systems to check alignment and safeguards over an extended period. Apollo Research will test whether models could covertly work against their developers' intentions. Goodfire will embed interpretability researchers, and METR will conduct two assessments.