OpenAI Dissolves Dedicated AI Risk Preparedness Unit

The responsibility for monitoring dangers, including biological and cyber threats, has been redistributed across existing engineering departments. Dylan Scandinaro, who moved to OpenAI from Anthropic only in February to lead the preparedness group, will transition to a new role focused specifically on the implications of recursive self-improving artificial intelligence.

This reorganization follows a period of heavy personnel turnover. Recent departures include chief futurist Josh Achiam, safety lead Johannes Heidecke, and ethics expert Chloé Bakalar. The trend has drawn sharp criticism from former staff, including Jan Leike, who resigned earlier this year citing a culture that prioritizes the release of shiny products over fundamental safety guardrails. By dissolving this team, the company continues its move away from the research-heavy structure that defined its early years, effectively folding long-term safety considerations into its standard operational workflow.