OpenAI supports President Donald Trump’s June 2 executive order establishing a voluntary framework for reviewing advanced AI models. The program invites developers to grant the federal government 30 days of access to new frontier models for security assessments. This initiative does not require mandatory licensing or pre-clearance.
CEO Sam Altman stated on X that the order maintains a proper balance. Altman is currently in Washington to discuss AI policy with administration officials and lawmakers.
OpenAI released a policy paper advocating for a civilian agency to oversee future mandatory evaluations. The current order grants the National Security Agency a lead role in these assessments. Google and Anthropic are also expected to participate in the voluntary collaboration.