President Donald Trump on Tuesday signed an executive order on artificial intelligence oversight, less than two weeks after delaying a White House ceremony over concerns his own policies could dull America’s technical edge.
The decree sets up a process for the federal government to examine the national security dangers of the most powerful AI systems for up to a month before they are released to the public. The ruling says participation by AI developers would be voluntary.
“The advanced AI capabilities that enhance our Nation also create new national security concerns that require coordinated action across executive departments and agencies,” the order states.
It wasn’t immediately obvious how the directive was different from one Trump refused to sign May 21.
The directive indicates the government would have just 30 days to assess an AI system, a shorter time frame than some in the business had anticipated. A lengthier time period could have been perceived as too cumbersome for a fast-moving, highly competitive business.
Trump scrapped a planned Oval Office meeting with tech industry executives last month because he was unhappy with the language in an earlier version of the order, the official said. “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 mandate was portrayed as a voluntary partnership with participating U.S.-based tech companies, including Anthropic, Google and OpenAI, which are commonly referred to as “frontier labs” since they are constructing the most advanced AI systems. Executives from several companies were slated to attend the signing event on May 21. Trump ultimately signed it without ceremony.
In a Tuesday social media post, the White House stated the executive order “establishes a process for frontier labs to voluntarily share cutting-edge cyber models to secure critical infrastructure and strengthen the government’s own cyber defenses. “We are NOT doing oversight of all new models, that level of government overreach would have chilling effects on free speech and innovation.
The directive is flawed, but “a step in the right direction to prepare the nation for the release of advanced AI systems,” said Juan Londoño, a policy analyst at the libertarian-leaning Cato Institute.
He praised the White House’s description 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.
Giving the NSA director that much latitude was a “dangerous precedent” that may enable the government to “weaponize” the policy against companies it is at odds with, such as Anthropic, Londoño said in an interview.
Plans for a new AI cybersecurity guideline came after Anthropic announced in April its most sophisticated AI model, named Claude Mythos, while the company is in a legal battle with the Trump administration over a contract disagreement with the Pentagon.
Soon after, Treasury Secretary Scott Bessent and outgoing Federal Reserve Chair Jerome Powell had an urgent conference with Wall Street CEOs, warning them of the hazards presented by Mythos’ apparent capacity to uncover cybersecurity weaknesses in the world’s software. Anthropic has restricted Mythos access to a small number of trusted partners, including large IT firms and banks, but on Tuesday it stated it had added another 150 organizations to that list.
Anthropic said Trump’s new order was “an important step in strengthening America’s leadership in AI” and added that it looked forward to working with the White House to promote its implementation.
Both Google and its main opponent, the maker of ChatGPT, OpenAI, said the policy was a significant step.
“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 Chris Lehane, chief global affairs officer at OpenAI, in a statement.
Democratic Sen. Mark Warner, vice chairman of the Senate Intelligence Committee, praised Trump’s policies but chastised the administration for having “belatedly discovered the need to redo something it hastily dismantled in its first year.”
Last year, just hours after returning to the White House, Trump tore down many of former President Joe Biden's guardrails for AI.
