President Trump on Tuesday signed an executive order asking developers of frontier artificial intelligence models to hand the federal government a 30-day cybersecurity review window before public release, a voluntary regime that arrived only after the White House quietly killed a stricter 90-day version less than two weeks earlier.

The order, titled “Promoting Advanced Artificial Intelligence Innovation and Security,” was signed without a ceremony. The previous draft had been scheduled for a public signing in May before the event was abruptly canceled under pressure from industry voices warning that a longer review would, in the administration’s own framing to NPR, blunt the American lead in competition with China. Trump told reporters he “didn’t like certain aspects of it.” The text he ultimately approved is careful to say so out loud: “Nothing in this section shall be construed to authorize the creation of a mandatory governmental licensing, preclearance, or permitting requirement for the development, publication, release, or distribution of new AI models, including frontier models.”

That sentence is the whole policy. Everything else is plumbing.

The plumbing is real. Within 30 days, the Secretary of Homeland Security must issue Binding Operational Directives covering civilian federal systems and extending CISA’s cybersecurity services to state and local authorities, rural hospitals, community banks, and local utilities. Within 60 days, the Office of Management and Budget has to identify federal grant programs redirectable toward A.I. vulnerability detection. Analysts at the Council on Foreign Relations describe the design as an attempt to “engineer a cybersecurity window of opportunity” through a classified NSA-run designation process, letting American defenders study a model before its capabilities reach adversaries. Whether that window can hold open against a technology that ships in months, they wrote, “remains to be seen.”

The order also lands inside an unresolved fight with the labs it depends on. In February, the Defense Department designated Anthropic a “supply-chain risk” after the company declined to waive its restrictions on mass surveillance and fully autonomous weapons. Anthropic is contesting the designation in court, and on Monday filed confidentially for an initial public offering.

A voluntary review depends on the goodwill of companies the administration is simultaneously suing and being sued by. The 30-day number is what that goodwill currently costs.

Sources