OpenAI announced on 28 May 2024 that its Board of Directors has established a Safety and Security Committee, a new governance structure intended to oversee critical safety and security decisions for the company's projects and operations [source].

The committee comprises four board members: Bret Taylor (Chair), Adam D'Angelo, Nicole Seligman, and Sam Altman. According to the announcement, the committee will evaluate and further develop OpenAI's processes and safeguards over a 90-day period, after which it will provide recommendations to the full board. The board will then review these recommendations and determine what information to share publicly.

OpenAI stated the committee will assist the full board in overseeing safety and security efforts across all company projects, including the development and deployment of artificial intelligence models. The announcement did not specify what prompted the formation of the committee or detail what existing oversight mechanisms it supplements or replaces.

The company described the move as part of its commitment to ensuring that artificial general intelligence benefits all of humanity. No timeline was provided for when the committee's recommendations would be made public, nor what specific safety or security concerns the committee would prioritise during its initial review period.

The formation follows OpenAI's ongoing deployment of GPT-4 and related models to millions of users. The announcement did not reference any specific incidents, model behaviours, or external pressure that led to the committee's creation. OpenAI indicated that further updates would be shared as the committee completes its work.