OpenAI Shuts Down AGI Safety Team as Key Leader Departs for New Venture

OpenAI has disbanded its AGI Readiness team, the group responsible for assessing the company’s preparedness for advancing AI technologies. Miles Brundage, the team’s lead, announced his departure this week, revealing plans to focus on independent AI policy research through a new or existing nonprofit. The disbanded team’s members will transition to other teams, with an economic research sub-group moving under OpenAI’s chief economist, Ronnie Chatterji.

Brundage expressed doubts about the readiness of AI labs, including OpenAI, for artificial general intelligence (AGI), defined as AI with human-level cognition. He shared concerns about AI’s safety and societal impact, stating, “I think AI is unlikely to be as safe and beneficial as possible without a concerted effort.” He emphasized his desire to contribute to AI policy across the industry, publishing more freely and independently.

OpenAI acknowledged Brundage’s exit, lauding his contributions to AI policy. “His decision to pursue policy research outside the industry gives him the opportunity to impact on a wider scale,” OpenAI said in a statement. The move follows OpenAI’s earlier disbanding of its superalignment team, initially formed to focus on AI’s existential risks, and the recent departure of several high-ranking leaders, including Chief Technology Officer Mira Murati.

The restructuring comes amid a period of significant change at OpenAI. After high-profile departures, the company completed a $6.6 billion funding round, achieving a valuation of $157 billion and reportedly considering changes to its nonprofit structure, potentially granting CEO Sam Altman substantial equity in the organization.