According to Odaily Planet Daily, OpenAI said on Monday that its CEO Sam Altman will leave the board's safety and security committee, which will now be composed entirely of independent board members.
The committee, which will be chaired by Zico Kolter, will be briefed on major models and, along with the full board, have the power to delay the release of new models, OpenAI said.
It noted that the new committee reviewed the safety of the recent o1 "Strawberry" model, which was rated as "medium risk" in the company's internal assessment.
OpenAI also said it is working to adopt the committee’s recommendations to increase transparency and collaboration with outside groups and to unify its safety practices across a growing number of product teams.