President Trump signed an executive order on June 2 establishing a voluntary federal framework for reviewing advanced AI models before public release. The order covers covered frontier models primarily those with advanced cyber capabilities and allows up to 30 days for national security review. The White House explicitly ruled out mandatory licensing or preclearance requirements.
OpenAI announced on June 3 that it would voluntarily comply with the framework, becoming the first major developer to commit to the review process. The company said it would submit future frontier models for federal assessment before release, aligning with the administrationcs national security priorities.
The order followed weeks of internal administration debate over the balance between security oversight and maintaining U.S. competitiveness in AI development. Industry analysts noted that voluntary compliance could slow release timelines for cutting-edge models, while the administration emphasized the need to guard against cyber threats posed by advanced AI systems.
The framework directs agencies to establish benchmarking standards and create a cybersecurity clearinghouse within 60 days. The voluntary structure means developers retain final authority over release decisions, a departure from earlier proposals that called for mandatory preclearance.
The market currently prices a 12% chance of a mandatory review order by June 30, reflecting trader uncertainty about whether additional executive action will convert the voluntary framework into a binding requirement before the deadline. No further White House announcements are scheduled.
