OpenAI revealed on November 16, 2026, that it will engage four research labs—METR, Apollo Research, Transluce, and Goodfire—to assess its AI models and safety features. However, there are concerns about the limited access these assessors will have to OpenAI's technology, which may compromise the independence of the evaluations. Previous assessments have highlighted issues with AI outputs, and critics argue that OpenAI's control over the process could hinder transparency. Additionally, OpenAI is facing a lawsuit from the Center for Investigative Reporting for alleged copyright violations.
Key facts
- OpenAI announced assessments by four labs on November 16, 2026.
- The assessors are METR, Apollo Research, Transluce, and Goodfire.
- Concerns exist regarding the limited access assessors will have to OpenAI's technology.
- Previous investigations indicated challenges in understanding AI outputs.
- The Center for Investigative Reporting has sued OpenAI for copyright violations.
A short AI summary in our own words. The reporting belongs to Mother Jones; we don't reproduce it. You can compare how outlets framed this story below.