OpenAI’s Frontier Governance Framework

OpenAI has announced the publication of its Frontier Governance Framework, detailing how the company's safety, security, and risk practices align with emerging legal requirements. According to OpenAI, the new framework specifically addresses compliance with California’s Transparency in Frontier AI Act and the EU AI Act’s Code of Practice for General Purpose AI. The foundation of this new governance document remains OpenAI’s Preparedness Framework, which outlines the company's internal approach to managing serious risks from advanced AI systems and includes practices that exceed current legal requirements. The Frontier Governance Framework applies relevant portions of that internal approach to specific public regulatory obligations. The document covers multiple risk assessment and mitigation areas, including cyber offense, chemical, biological, radiological, and nuclear risks, harmful manipulation, and loss of control. Additionally, it addresses model reporting, security risk management, incident response, external expert input, and procedures for framework updates. OpenAI stated that it expects its approach to continue evolving as model capabilities, evaluations, and regulatory requirements develop, and the company intends to update the Frontier Governance Framework accordingly as the landscape changes.