OpenAI Disbands Preparedness Team Amid Safety Restructuring

OpenAI disbanded its Preparedness team in July 2026, a group previously responsible for assessing catastrophic risks including biological hazards, cybersecurity threats, and autonomous replication. The company is redistributing these responsibilities across existing specialized divisions to integrate safety work more closely with model development.

While the Financial Times reported the team was shut down, spokesperson Kayla Wood denied the claim, stating that research leaders in cybersecurity, biological and chemical, and AI self-improvement capabilities continue to report to head of safety Saachi Jain. Despite the reorganization, OpenAI maintains that its Preparedness Framework remains the foundation for managing serious risks.

This move follows the previous dissolution of the superalignment and AGI readiness teams, as well as the departure of safety systems head Johannes Heidecke and ethics lead Chloé Bakalar. Former employee Jan Leike criticized the company for prioritizing "shiny products" over safety. Dylan Scandinaro, the former leader of the Preparedness team, has shifted his research focus toward recursive self-improving AI. The restructuring occurs as the company prepares for a massive IPO, leading critics to suggest that commercial speed is being prioritized over independent safety oversight.