President Donald Trump signed an executive order Tuesday directing federal agencies to establish a voluntary framework for reviewing the country’s most powerful artificial intelligence systems before public release, with the National Security Agency (NSA) assigned a central role in identifying which models warrant government scrutiny.
The order asks leading AI developers, including Anthropic, OpenAI, and Google, to voluntarily provide federal access to advanced systems up to 30 days before they are deployed to other trusted partners. The framework relies on classified benchmarking administered by the NSA to determine whether a system qualifies as a “covered frontier model” based on its national security and cyber risks.
Led by the Treasury Department, the clearinghouse will operate in consultation with the NSA and the Cybersecurity and Infrastructure Security Agency (CISA). This collaborative body will serve as a centralized hub to coordinate vulnerability scanning, validate software flaws, and prioritize the rollout of security patches.
“AI systems are now the most powerful tools we have ever had to harden our cyber infrastructure and stay ahead of adversaries,” Michael Kratsios, director of the White House Office of Science and Technology Policy (OSTP), wrote on X. “It is a blessing that these capabilities are being developed by American industry, and not by those who would use them against us.”
The order bars the government from imposing mandatory licensing or preclearance requirements on AI developers. Trump had delayed signing an earlier version in May over concern that a 90-day review window would disadvantage U.S. companies against Chinese competitors. The final order cut that window to 30 days.
Anthropic’s Mythos Preview model, released confidentially in April with a demonstrated capability to detect critical software vulnerabilities, drove much of the urgency. Anthropic has since expanded its Project Glasswing access program from about 50 organizations to nearly 200 across more than 15 countries.
The Department of Defense (DOD) previously designated Anthropic a supply chain risk, prohibiting defense contractors from using its technology. That litigation remains pending in Washington, D.C.
