OpenAI creates security committee as it starts training new model

(Reuters) -OpenAI has formed a Safety and Security Committee that will be led by CEO Sam Altman as it begins training its next artificial intelligence model, the AI ​​startup said on Tuesday.

Directors Bret Taylor, Adam D’Angelo and Nicole Seligman will also lead the committee, OpenAI said in a company blog.

Former chief scientists Ilya Sutskever and Jan Leike, who were leaders of the Microsoft-backed OpenAI Superalignment team that ensured AI remained aligned with its intended goals, left the company earlier this month.

OpenAI disbanded the Superalignment team in early May, less than a year after the company created it, with some team members reassigned to other groups, CNBC reported days after the high-profile departures.

The committee will be responsible for making recommendations to the board regarding security decisions for OpenAI projects and operations.

His first task will be to evaluate and further develop OpenAI’s existing security practices over the next 90 days, after which he will share recommendations with the board.

Following the board’s review, OpenAI will publicly share an update on the adopted recommendations, the company said.

Other committee members include the company’s technical and policy experts Aleksander Madry, Lilian Weng and head of alignment science John Schulman. Newly appointed chief scientist Jakub Pachocki and security chief Matt Knight will also join the committee.

(Reporting by Arsheeya Bajwa and Akash Sriram in Bengaluru; Editing by Tasim Zahid and Vijay Kishore)