Bottom line: OpenAI has disbanded its standalone Preparedness team for assessing catastrophic AI risks and distributed its responsibilities across existing teams.
OpenAI has disbanded its Preparedness team, which was responsible for examining catastrophic risks of AI models. Its previous responsibilities are now to be taken over by existing teams.
The Preparedness team at OpenAI was responsible for examining AI models for potentially catastrophic risks. This included scenarios in which models could be misused for the development of biological, chemical, radiological or nuclear weapons, as well as risks in the areas of cybersecurity and autonomous AI systems. According to a report by Heise, this standalone team has now been disbanded, with its responsibilities distributed among other, already existing teams within the company.
For security officers at companies, this organizational change is relevant because it raises questions about the internal prioritization of risk assessments at one of the leading AI providers. When specialized teams for assessing catastrophic risks are disbanded and folded into more generalist structures, this can affect the depth and consistency of future risk analyses — particularly for new model generations deployed in enterprise environments.
In practice, this means that companies deploying OpenAI models in critical processes should continue to carefully review the publicly available safety and risk assessments (such as system cards). It is also advisable to monitor how the distribution of Preparedness tasks to other teams specifically affects the transparency and documentation of future model releases. From a regulatory perspective, this development could also become relevant in the context of ongoing debates about AI governance and providers’ oversight obligations.
Source: www.heise.de · Published August 17, 2026
Lumi AI News — AI-assisted curation in accordance with Art. 50 EU AI Act. Paraphrasing and classification by Lumi News Pipeline v1.8.3.