OpenAI launches ‘independent’ safety board that could halt release of its models

OpenAI Transforms Its Safety and Security Committee into an Independent “Board Oversight Committee” with the Power to Delay Model Launches Due to Safety Concerns according to the OpenAI blog postThe commission recommended the creation of the independent board following a recent 90-day review of OpenAI’s “safety and security processes and safeguards.”

The committee, chaired by Zico Kolter and comprised of Adam D’Angelo, Paul Nakasone, and Nicole Seligman, “will be briefed by company management on security assessments for major model releases and will, in conjunction with the full board, provide oversight for the release of models to market, including the authority to delay release until security issues are resolved,” OpenAI says. OpenAI’s full board of directors will also receive “periodic briefings” on “security issues.”

OpenAI’s security committee members are also members of the company’s broader board, so it’s unclear how independent the committee is or how that independence is structured. (CEO Sam Altman previously served on the committee but no longer does.) We’ve reached out to OpenAI for comment.

By establishing an independent security board, OpenAI appears to be taking an approach similar to the Meta Oversight Board, which reviews some of Meta’s content policy decisions and can make rulings that Meta must follow. None of Members of the Supervisory Board are on Meta Board.

The review by the OpenAI Safety and Security Committee also helped to provide “additional opportunities for industry collaboration and information sharing to advance the security of the AI ​​industry.” The company also says it will seek “more ways to share and explain our security work” and “more opportunities for independent testing of our systems.”

Update, September 16:Added that Sam Altman is no longer a member of the committee.