The White House introduced a voluntary review framework for frontier AI models following an executive order in June 2026. This system aims to assess cybersecurity risks before advanced models are released to partners, shifting control from developers to national security authorities.
OpenAI and Anthropic have already begun restricting access to their latest systems. OpenAI plans to limit its GPT-5.6 Sol model to government-approved partners, while Anthropic is restricting its Mythos 5 model to a select group of cyber defense professionals.
In July 2026, the administration launched the Gold Eagle AI cybersecurity clearinghouse to centralize AI-driven defenses for critical infrastructure. However, as of mid-September 2026, the government has not publicly disclosed the specific qualification criteria or reporting standards.
The lack of clarity has created uncertainty for both domestic firms and international allies. US officials and G7 representatives have been discussing access protocols since mid-June 2026, as other nations seek to secure their own pathways to these advanced technologies.
