OpenAI disbanded its preparedness team, the group that assessed whether its models pose catastrophic risks, at the end of July, the Financial Times reported. Engadget and Digit.in both confirmed the FT's reporting.
The team evaluated risks such as models being used to develop biological or cyber weapons, or acting autonomously in dangerous ways, and built mitigation strategies for those scenarios, according to Digit.in. OpenAI has split those duties among senior staff in other divisions, assigning specific people to areas like bio and cyber risk rather than keeping one dedicated team; the team's former lead, Dylan Scandinaro, now focuses on risks from recursive self-improving AI, Digit.in reported.
OpenAI described the move as part of a "streamlining process" ahead of its planned initial public offering, after chief executive Sam Altman asked employees to cut back on side projects and focus on ChatGPT, Engadget reported. The company also shut down its Sora video generation product around the same time.
The reorganization follows recent departures of ethics lead Chloe Bakalar, chief futurist Josh Achiam and safety chief Johannes Heidecke, according to Digit.in. It's the third safety-focused unit OpenAI has dissolved in about two years, after the AGI Readiness team in 2024 and the Mission Alignment team earlier in 2026, per Engadget.
For developers building on OpenAI's models, the practical read is narrower than the headlines: risk assessment didn't disappear, it moved into existing teams. Whether that holds up gets tested the next time a model does something its builders didn't expect.