1 hour ago
9
OpenAI quietly disbanded its Preparedness team at the end of July, according to the Financial Times. The team’s entire mandate was evaluating whether frontier AI models posed serious risks, things like rogue systems hacking other companies or enabling biological threats, and then figuring out how to stop those scenarios from happening.
Instead of maintaining a dedicated unit, OpenAI has carved up those responsibilities by domain, splitting bio and cyber risk assessment across existing teams.
A pattern, not an incident
This would be notable on its own. But it’s the third safety-focused team OpenAI has dissolved in roughly two years, which starts to look less like reorganization and more like a trend.
The AGI Readiness team was disbanded in 2024. The Mission Alignment team was closed in February 2026. Now the Preparedness team joins them.
Each dissolution came with a version of the same explanation: the work isn’t going away, it’s just being integrated into other parts of the organization. Co-founder Greg Brockman has suggested this approach actually strengthens safety by embedding it directly into model development and product teams rather than siloing it in a separate group.
Executive departures compound the concern
The team dissolution hasn’t happened in a vacuum. OpenAI has also seen significant executive turnover that touches directly on safety and governance.
COO Brad Lightcap and Chief Ethics Officer Chloé Bakalar have both departed. Dylan Scandinaro, who formerly headed the Preparedness team, has reportedly shifted his focus to risks from self-improving AI systems, but the team he built no longer exists as a unit around him.
Internal sources have flagged concerns about what these repeated departures mean for safety work. When the people who built institutional knowledge around risk assessment leave or get reorganized, continuity suffers.
The IPO elephant in the room
All of this is happening as OpenAI prepares for what is expected to be one of the largest tech IPOs in years. The company has been steadily moving away from its original research-led nonprofit structure toward a more conventional corporate model, a transformation that has been years in the making.
OpenAI has maintained a Preparedness Framework since at least 2024, aimed at tracking and mitigating severe harms from advanced AI technologies. The framework was updated as recently as April 2025 to define capability thresholds of High and Critical, along with the introduction of Safeguards Reports reviewed by an internal Safety Advisory Group.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·