OpenAI Dissolves AI Preparedness Team to Prioritize Product Growth

OpenAI has reportedly dissolved its preparedness team, a group that evaluated the risks posed by its language models, according to the Financial Times. The team was shut down at the end of last month, after the company announced a broader restructuring of its policy and safety functions. The move comes as OpenAI prepares to launch new products and expand its user base.

The decision matters because AI safety has become a headline issue. High‑profile incidents, such as the GPT‑4 jailbreaks and the spread of disinformation, have pushed regulators and the public to demand tighter controls. OpenAI’s own policy team had been responsible for spotting potential misuse, setting usage guidelines, and designing technical safeguards. By removing this internal safety watchdog, the company risks losing a key line of defense against harmful applications of its models.

There is no public statement from OpenAI explaining the rationale for disbanding the team. Industry observers note that the company has been trimming costs and shifting resources toward product development. Analysts point out that the European Union’s AI Act and U.S. congressional inquiries could pressure OpenAI to demonstrate robust safety protocols. Some AI safety experts warn that without a dedicated team, the company may be slower to identify new risks or to update its mitigation strategies.

What happens next is unclear. OpenAI may replace the preparedness group with a smaller compliance office or outsource safety reviews to external auditors. If it does not, regulators could step in to require independent oversight. The company’s next public move will likely involve a new safety roadmap or a partnership with an academic research lab. The outcome will shape how quickly OpenAI can roll out new models while keeping safety at the forefront.