Trump Signs a Softer AI Executive Order: A Voluntary 30-Day Frontier-Model Review, and No Mandatory Licensing

President Donald Trump on June 2 signed an executive order titled “Promoting Advanced Artificial Intelligence Innovation and Security,” asking the country’s leading AI developers to voluntarily give the federal government access to their most capable models for up to 30 days before they ship to the public. The goal, the order says, is to let agencies run a classified benchmarking process that assesses a model’s “advanced cyber capabilities” and decide whether it should be treated as a “covered frontier model.” The companies in the frame are the usual frontier trio — Anthropic, OpenAI and Google DeepMind — alongside other labs building at the frontier.

What is most notable is what the order does not do. Rather than impose pre-deployment testing as a condition of release, it leans entirely on goodwill, and it says so in black and white: “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.” The review framework is to be coordinated through the National Institute of Standards and Technology, with developers free to engage agencies to determine whether a given model qualifies and to help select the “trusted partners” that get early access.

The 30-day window is itself the product of a fight inside the administration. An earlier draft circulated in May contemplated a 90-day voluntary review — long enough, labs warned, to delay product launches. The administration’s AI and crypto czar, David Sacks, pushed to cut it to 30 days and to strip out anything mandatory. “The change in the EO from a 90 day to 30 day period is a game changer because it allows our AI labs to comply with the voluntary framework without delaying new model releases,” Sacks said.

The order’s second half is squarely about defense. It directs the Treasury Department to stand up an “AI cybersecurity clearinghouse” to coordinate the use of advanced AI for software vulnerability scanning, discovery, validation and patching, and it tasks the Cybersecurity and Infrastructure Security Agency with helping agencies, state and local authorities, and operators of critical infrastructure — “rural hospitals, community banks, and local utilities” among them — get their hands on those tools, including covered frontier models where appropriate.

For Trump, the signing is also a reversal of a reversal. A tougher AI safety order was scheduled to land in late May before being pulled at the last minute following calls from industry figures including Elon Musk, Mark Zuckerberg and Sacks. The version that actually made it onto paper is a far lighter-touch document: innovation-first in its framing, voluntary in its mechanism, and careful to avoid the licensing regime that Silicon Valley had lobbied hard against.

Reaction split predictably along the safety-versus-speed divide. Cato Institute analyst Juan Londoño called the order “imperfect” but “a step in the right direction,” while warning that the lack of clear criteria for what counts as a covered frontier model — and the government’s hand in choosing trusted partners — “gives the executive a great deal of discretion” that “could open the door to potential weaponization against companies that have any sort of conflict with the administration.” Former FTC chief technologist Neil Chilson said the order beats the current informal approach but urged Congress to set actual rules, and the Center for Democracy and Technology’s Samir Jain said his group would be “closely monitoring” implementation so the order does not “become a mechanism for the Administration to punish companies for political or other arbitrary reasons.” The worry is not abstract: the Pentagon recently labeled Anthropic a supply-chain risk, and the company has sued to reverse the designation.