OpenAI disbands preparedness team amid broader safety team shakeups

OpenAI disbanded its preparedness team at the end of July 2026, according to a report by the Financial Times. The team’s role was to assess whether models posed serious risks and to develop mitigation strategies.

Responsibility for risk assessment has been divided among existing teams, with specific domains such as bio and cyber now handled separately.

The move is the latest in a series of changes to OpenAI’s safety structure, following the dissolution of its AGI readiness and superalignment teams in recent years.

Recent departures from OpenAI’s safety organization include ethics lead Chloé Bakalar, Chief Futurist Josh Achiam, and head of safety Johannes Heidecke.

Former preparedness team head Dylan Scandinaro, who joined OpenAI from Anthropic in February 2026, will now focus on the implications of recursive self-improving AI.

Criticism of the changes has come from former OpenAI researcher Jan Leike, who resigned in 2024 and told the Financial Times that the company was prioritizing product development over safety.