OpenAI’s Frontier Governance Framework

OpenAI has published a Frontier Governance Framework that describes how its safety and security practices meet new legal requirements from California and the EU, building on its existing Preparedness Framework for managing risks from advanced AI systems. The framework covers risk assessment and mitigation in areas like cyber attacks, dangerous biological/chemical/nuclear risks, manipulation, and loss of control, along with model reporting and incident response. OpenAI says it will update this framework as AI capabilities and regulations evolve.