Trump signs executive order for voluntary AI vetting on national security risks

President Donald Trump signed an executive order Tuesday establishing a voluntary 30-day federal review of advanced artificial intelligence systems before their public release, citing national security concerns raised by rapidly advancing AI capabilities.

The order marks the first White House framework for vetting AI models before they reach consumers — arriving as Anthropic's Claude Mythos model has already prompted urgent cybersecurity discussions among federal officials, according to ABC7 Los Angeles.

The order tasks the director of the National Security Agency with overseeing evaluations of AI models that developers voluntarily submit for review. The government must complete its assessment within 30 days before a system can be publicly released. Participation by AI developers — including those from Anthropic, OpenAI and Google — is voluntary, ABC7 Los Angeles reported.

The signing came less than two weeks after Trump abruptly called off a similar ceremony set for May 21. The president had expressed concern that the original policy could blunt America's technological edge over China. "We're leading China, we're leading everybody, and I don't want to do anything that's going to get in the way of that lead," Trump told reporters at the time.

The final order frames AI as both an asset and a risk. "Advanced AI capabilities make our Nation stronger, but also introduce new national security considerations that require coordinated action across executive departments and agencies," the order text states, according to ABC7.

The ABC7 report noted that Anthropic's Claude Mythos model prompted urgent discussions among Treasury Department and Federal Reserve officials about cybersecurity vulnerabilities following its announcement. Anthropic subsequently expanded access to Mythos for 150 additional organizations.

Juan Londoño, a policy analyst at the Cato Institute, described the order as imperfect but "a step in the right direction to prepare the nation for the release of advanced AI systems." In a written statement, Londoño also cautioned that vague criteria for model selection and a "trusted partner" designation carried serious risks. "This could open the door to potential weaponization against companies that have any sort of conflict with the administration," he said.

The White House has not clarified which AI models qualify for review or what criteria govern the "trusted partner" designation, leaving policy observers to warn that the framework's real impact — whether a genuine national security safeguard or a mechanism for political pressure — will hinge on those still-undefined details.