OpenAI dissolves preparedness team ahead of IPO

OpenAI dissolved its preparedness team at the end of July, folding oversight of bio, cyber and rogue-system risk into existing teams, per the Financial Times. The team's job was assessing whether frontier models posed serious harm before release. The Verge reported the dissolution over the weekend, along with OpenAI's framing: a streamlining move ahead of the company's expected IPO.

It's the third safety-focused group OpenAI has shut in about two years. AGI Readiness closed in 2024. Mission Alignment closed this February. Ethics lead Chloé Bakalar and safety head Johannes Heidecke have both left the company since.

Altman told staff to cut "side quests" and focus on core ChatGPT work. Distributing preparedness's framework domains to existing teams fits that instruction: fewer standalone groups, fewer headcount lines that don't map to product. It also removes a team whose entire mandate was to say no.

Distributed ownership of a risk framework is functionally different from a team whose job it is. When bio-risk assessment sits inside a product team with a deadline, flagging a delay competes with hitting it. A dedicated team without that conflict was the buffer a pre-IPO OpenAI removed. The Financial Times reported the dissolution over the weekend, tying it to OpenAI's move toward an IPO. Whether the new owners get the authority, or the willingness, to slow a launch isn't something either report says.