The past month has seen significant changes at OpenAI, with several high-ranking executives departing, advocating for increased control over the spread of artificial intelligence technologies. Alongside these departures, a special unit within the company was disbanded. However, its functions will largely be transferred to a newly created supervisory committee on the board of directors.
Formation and Responsibilities of the Oversight Committee
As reported by Bloomberg, the new committee will include both board members and company executives. Its primary role will be to assess the security of large language models developed by OpenAI. The committee has been given 90 days to conduct an initial assessment of existing safety regulations. After this period, the committee will publish its recommendations and make them public. Additionally, OpenAI has announced the commencement of training its latest large language model, most likely GPT-5. We’ll keep you updated on this development.
Impact on OpenAI’s Internal Structure
Previously, OpenAI’s specialized division struggled for resources, leading to internal conflicts. Leading scientific specialist Ilya Sutskever, who left the company in May, had disputes over risk assessment methods with current management last fall. This conflict resulted in an attempt to resign the company’s founder, Sam Altman. Part of the responsibility for monitoring the societal impact of artificial intelligence has been assigned to the OpenAI research division. Following Sutskever’s departure, John Schulman, one of the startup’s original members, has taken over this division.
Composition of the Oversight Committee
The oversight committee will consist of three board members: Chairman Bret Taylor, CEO Adam D’Angelo, and former Sony Entertainment executive Nicole Seligman. Additionally, six committee members will be drawn from the company’s executives, including Altman and Schulman. The committee will also involve third-party advisers, including two former high-ranking U.S. officials, adds NIXSolutions. These measures reflect OpenAI’s commitment to enhancing the safety and transparency of its AI technologies.
In summary, OpenAI is undergoing a period of significant restructuring aimed at improving the oversight and safety of its AI developments. The formation of the new supervisory committee and the training of GPT-5 are key steps in this process. We’ll keep you updated on further developments.