OpenAI disbanded its preparedness team at the end of July, The Verge reported.
Why it matters: The team was part of OpenAI's process for reviewing frontier-model risks, so its removal could affect how severe safety issues are assessed before launches. It also leaves open whether that review work has been centralized elsewhere or reorganized without the same public structure.
- The Verge reported Aug. 16, citing the Financial Times, that OpenAI disbanded its preparedness team at the end of July.
- OpenAI created the preparedness function in 2023 to track frontier risks including cybersecurity, CBRN threats, autonomous replication and individualized persuasion.
- OpenAI said May 28 that its Preparedness Framework remained the basis for how it operationalizes serious AI risk management.
- OpenAI's governance framework says the Safety Advisory Group reviews safeguards and makes recommendations to leadership.
OpenAI's preparedness team was created in 2023 to help identify, track and prepare for frontier AI risks, including cybersecurity, CBRN threats, autonomous replication and individualized persuasion, according to the company’s launch post. OpenAI said the team would connect capability assessment, evaluations and internal red teaming for frontier models.
The Verge reported Aug. 16, citing the Financial Times, that OpenAI disbanded the team at the end of July. The Financial Times story was not available in the dossier, but The Verge said the change happened last month. The report comes amid a broader reshuffling of safety-related work at OpenAI.
OpenAI's public governance materials still describe a formal risk-review structure. In a May 28 post, the company said the Preparedness Framework remained the foundation for how it operationalizes its approach to the most serious AI risks. OpenAI's governance framework also says the Safety Advisory Group reviews whether safeguards sufficiently minimize severe risk and makes targeted recommendations.
The available sources do not identify the team’s successor structure, if any. They also do not say whether the functions formerly handled by preparedness staff were moved to another group or folded into other internal processes.
By the numbers
- 2023 - OpenAI created the preparedness function to track frontier risks.
- May 28 - OpenAI said its Preparedness Framework remained the basis for serious AI risk management.
- End of July - The Verge reported the team was disbanded then.
Yes, but: The sources do not confirm whether another team absorbed the preparedness function or what replaced it.