OpenAI dissolves dedicated safety team, scatters risk work across groups
OpenAI dismantled its dedicated Preparedness team, which conducted internal evaluations of catastrophic AI risks from the company's own models. The safety function has been distributed across existing departments, triggering staff departures and internal concern about fragmented oversight. This structural shift signals a potential deprioritization of systematic risk assessment at a leading frontier lab, raising questions about whether distributed safety responsibilities can match the rigor of a focused team during a period of rapid capability scaling.
Modelwire context
Analyst takeThe dissolution isn't just a reorg on paper: the staff departures suggest the people who held institutional knowledge about evaluation methodology are leaving, not transferring, which means the distributed model loses the human continuity that made the original team coherent.
This is largely disconnected from recent activity in our archive, as we have no prior coverage to anchor it to. It belongs to a longer pattern visible across the frontier lab space, where dedicated safety or alignment teams have faced repeated structural pressure as compute investment and deployment timelines accelerate. The Preparedness team was itself a relatively recent creation, formed partly in response to public and regulatory scrutiny after GPT-4. Dismantling it now, during a period when capability scaling is faster than at any prior point, inverts the original logic: the team existed precisely because rapid scaling was the threat. Distributing that function across groups with competing product priorities is a structural bet that coordination costs won't outweigh the loss of focus.
Watch whether OpenAI publishes a Preparedness-style model evaluation report for its next major release under the new distributed structure. If no equivalent document appears within six months of the next flagship model launch, that is a concrete signal that the function was deprioritized rather than reorganized.
This analysis is generated by Modelwire’s editorial layer from our archive and the summary above. It is not a substitute for the original reporting. How we write it.
MentionsOpenAI · Preparedness team
Modelwire Editorial
This synthesis and analysis was prepared by the Modelwire editorial team. We use advanced language models to read, ground, and connect the day’s most significant AI developments, providing original strategic context that helps practitioners and leaders stay ahead of the frontier.
Modelwire summarizes, we don’t republish. The Decoder originally reported this story as “OpenAI dissolved the team built to catch catastrophic AI risks, reassigning its work to other groups”. The full content lives on the-decoder.com. If you’re a publisher and want a different summarization policy for your work, see our takedown page.