OpenAI Reportedly Disbands Its Preparedness Safety Team in Major AI Governance Shift

OpenAI has quietly disbanded its specialized Preparedness team—the internal unit formed to monitor and evaluate catastrophic risks associated with frontier AI models, including cybersecurity weapons and autonomous bioweapon synthesis.

Reorganizing Frontier Risk Evaluation into Product Safety Units

According to internal communications reported by The Verge, the decision transfers key safety researchers into broader product safety and alignment engineering groups. OpenAI leadership stated the consolidation will integrate safety checks directly into release engineering rather than maintaining a separate advisory body.

Industry Concerns Over Governance Oversight and External Audits

However, former safety advisors and industry observers have expressed concern that eliminating an independent risk-monitoring division weakens internal pushback against fast-paced commercial product launches.