WHITE HOUSE AI FRAMEWORK TO STAY PRIVATE

Three sources told Axios that the Biden administration plans to keep its new AI framework confidential rather than making it available to the public. The framework is designed to evaluate advanced AI models and will only be shared with companies involved in the voluntary process. The White House's approach specifically targets "frontier models"—defined as closed-source systems with state-of-the-art capabilities and potential national security implications. Open-source models are excluded from the framework's scope. The decision to restrict access reflects the administration's focus on managing risks from the most advanced proprietary AI systems, while allowing open-source development to proceed outside the regulatory framework. Companies working with the White House on AI safety commitments will receive framework details, but the general public and broader industry will not have access to the specific evaluation criteria and processes. This approach differs from typical government regulatory frameworks, which are typically published for public comment and transparency.