The order establishes a framework for the federal government to vet the national security risks of the most advanced AI systems for up to a month before their public release. Participation by AI developers would be voluntary, the order says.

“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 says.

It was not immediately clear to what extent the order differed from the one Trump declined to sign on May 21.

The order says the government would have only 30 days to review an AI system, a shorter time frame than some in the industry were expecting. A longer time period might have been seen as too burdensome for a fast-moving and highly competitive industry.

Trump canceled an Oval Office event with tech industry executives last month because he did not like what he saw in the earlier version of the order’s text. “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.

That directive was characterized as a voluntary collaboration with participating U.S.-based tech companies, including Anthropic, OpenAI and Google, which are sometimes described as “frontier labs” because they are building the most advanced AI systems. Several companies had been planning to have executives present at the May 21 signing event. Trump ended up signing it without any ceremony.

The White House said in a social media post Tuesday that the executive order “creates a process for frontier labs to voluntarily share cutting-edge cyber models in order to secure critical infrastructure and strengthen the government’s own cyber defenses. We are NOT conducting oversight of all new models, as that level of government overreach would have chilling effects on free speech and innovation.”

Juan Londoño, a policy analyst at the libertarian-leaning Cato Institute, said the order is imperfect but “a step in the right direction to prepare the nation for the release of advanced AI systems.”

He applauded the White House’s characterization of the process as voluntary but said he was concerned about the vagueness of how the government, led by the director of the National Security Agency, will decide which AI models qualify for scrutiny, and how it will decide which “trusted partners” get early access to them.

Londoño said in an interview that giving so much discretion to the NSA director was a “dangerous precedent” that could enable the government to “weaponize” the policy against companies it is clashing with, like Anthropic.

Plans for a new AI cybersecurity directive followed Anthropic’s April announcement of its most advanced AI model, called Claude Mythos, in the middle of the company’s legal fight with the Trump administration over a contract dispute with the Pentagon.

Treasury Secretary Scott Bessent and outgoing Federal Reserve Chair Jerome Powell soon after convened an urgent meeting with Wall Street CEOs, warning them about the risks posed by Mythos’ apparent ability to find cybersecurity vulnerabilities in the world’s software. Anthropic has limited access to Mythos to only a small group of trusted partners, such as big tech companies and banks, though it said Tuesday it has expanded that group by another 150 organizations.

Anthropic called Trump’s new order “an important step in strengthening America’s leadership in AI” and said it looks forward to collaborating with the White House to support its implementation.

Its chief rival, ChatGPT maker OpenAI, also described the policy as an important step, as did Google.

“As AI capabilities continue to advance, we believe effective safety frameworks should continue to be developed through democratic institutions, informed by technical expertise and broad stakeholder input, to promote accountability and public trust,” said a statement from Chris Lehane, OpenAI’s chief global affairs officer.

Democratic Sen. Mark Warner, vice chairman of the Senate Intelligence Committee, also welcomed Trump’s policy but criticized the administration for having “belatedly discovered the need to redo something it hastily dismantled in its first year.”

Trump repealed many of former President Joe Biden’s guardrails for AI just hours after returning to the White House last year.

—

AP Technology Writer Barbara Ortutay contributed.