OpenAI quietly axes its safety watchdog unit ahead of blockbuster IPO push

OpenAI shut down its preparedness team at the end of April, according to a report from the Financial Times. The team's core mandate was to evaluate whether OpenAI's models presented serious risks to society and to develop mitigation strategies before those models reached the public. Its dissolution marks a notable structural shift at a company that has long positioned safety as central to its identity.

Rather than keeping those responsibilities under one roof, OpenAI has reportedly parceled them out across existing internal teams organized by threat category — with separate groups now handling concerns in areas like biosecurity and cyber threats. On paper, that kind of specialization could mean deeper domain expertise. Critics would argue, however, that consolidating safety evaluation under a single independent team existed precisely to provide oversight that crosses departmental lines and avoids conflicts with product timelines.

The change is the latest in a string of organizational shifts at OpenAI as the company moves toward what observers expect to be one of the largest IPOs in tech history. Over the past few years, OpenAI has steadily unwound aspects of its original nonprofit-first governance structure, gradually transitioning toward a more conventional for-profit model. The restructuring has come alongside significant executive departures, including several figures who were publicly associated with the company's safety-forward positioning.

Founded in 2015 with explicit commitments to responsible AI development, OpenAI built its early brand on the promise that it would not race to deploy systems faster than it could ensure they were safe. The preparedness team was seen as an institutional expression of that commitment. Its disbanding raises questions about whether organizational accountability for frontier risk assessment has genuinely been preserved or simply made less visible.

The timing is particularly notable given that regulators in both the United States and European Union are actively scrutinizing AI labs' internal safety practices. Any perception that commercial pressures are crowding out safety infrastructure could invite closer regulatory attention precisely when OpenAI is trying to tell a compelling growth story to prospective investors.