OpenAI released its Frontier Governance Framework today, outlining how the company's safety and security practices map to emerging AI regulations. The document specifically addresses compliance with California's Transparency in Frontier AI Act and the EU AI Act's Code of Practice for General Purpose AI, translating internal protocols into publicly accountable governance structures.

The framework builds on OpenAI's existing Preparedness Framework, which guides risk management for advanced AI systems. While the Preparedness Framework covers internal practices exceeding current legal requirements, the Frontier Governance Framework focuses on meeting specific regulatory obligations through public documentation and accountability measures.

Risk assessment and mitigation cover cyber offense capabilities, CBRN threats, harmful manipulation, and loss of control scenarios. The framework also establishes protocols for model reporting, security risk management, incident response procedures, and incorporates external expert input into governance decisions.

OpenAI plans to evolve this approach as model capabilities and regulatory requirements advance, with updates expected through Framework2026. The company's proactive documentation signals industry recognition that AI governance must keep pace with rapid technological development.