Trump signs an executive order to vet top AI models for national security risks

WASHINGTON— President Trump signed administrative order The federal government was ordered Tuesday to establish a voluntary early review process for the nation’s most advanced artificial intelligence models. A civil war that lasted for months on how aggressively Washington should move to regulate the fast-growing technology.

Under the order, companies are asked to allow government agencies, including representatives from the National Security Agency and the Department of Defense, to evaluate the latest models up to 30 days before they are released to the public. The decision stops short of mandating opt-in and expressly prohibits the creation of any new licenses or permits for AI models.

“The question is whether this is the beginning of an ongoing crackdown by the government and a response to ongoing AI capabilities, or whether this is a one-time, limited, truly voluntary action,” said James Sanders, a research fellow at the Center for a New American Security, a think tank in Washington, D.C.

“As AI labs try to maintain good relations with the U.S. government, it is unclear how voluntary this will remain and how voluntary it will be in practice,” he said.

This order represents a reversal for Trump in less than two weeks. soaked A version of the policy that gives the government a 90-day review period — and more broadly, promises to remove AI guardrails for an incoming administration, a stance that is slowly creating rifts within the GOP.

In the executive order, Trump appeared to frame the need to develop artificial intelligence technologies while considering national security. “As these capabilities develop, my Administration will continue to work closely with industry to ensure the rapid deployment of the best and safest technology to counter any threats to our country,” the order said.

This step immediately sparked debate about whether Trump’s plan would be an effective approach. This formalizes an existing practice where top AI companies share models with external evaluators and government actors before making them available to the public, but it raises questions about how voluntary this will be and how the government will decide which labs to target.

David Sacks, who previously served as Trump’s AI adviser, called the 30-day period a “game changer” and argued that the shorter timeline would allow companies to engage with the government without slowing down new model releases.

“In the AI ​​race, every day counts,” Sacks wrote in a post on X.

Dean W. Ball, Trump’s former AI adviser, called the order a victory for the AI ​​”security alliance” and a loss for Sacks and others who support a more expedited approach. He called the order a mistake and said it could be the first step toward requiring federal licensing for artificial intelligence models.

“All for the benefit that can barely be articulated; what exactly will the intelligence community do in 30 days to make models safer?” Ball wrote to X.

The signing of the executive order comes amid growing tensions among Republicans over artificial intelligence, job loss and data center construction, including fears among a significant number of Trump’s supporters that artificial intelligence could eliminate jobs or become a security threat. Surveys conducted in May showed strong support There has been discussion among Republicans for a framework similar to the one outlined in Trump’s executive order.

The growing divide among Republicans on artificial intelligence was clearly visible in Florida on Monday; said the state’s Republican attorney general, James Uthmeier. A lawsuit was filed against OpenAI On ChatGPT’s alleged risks, an armed attacker used the bot at Florida State University last year.

Meanwhile, Trump’s endorsed candidate to replace Governor Ron DeSantis is Rep. Byron Donalds – said monday He stated that he disagrees with Trump on AI policy and supports government-led regulation; It was supported by the AI ​​industry earlier in the year.

A. questionnaire The report, released by Americans for Responsible Innovation, a nonprofit organization that advocates for a federal framework for AI policy, found that a majority of Republican voters surveyed support the type of plan laid out in Trump’s executive order. Seventy-one percent also said independent security testing for advanced AI systems should be required by law.

When Trump took office, his administration reversed the US stance on regulation, moving away from Biden-era policies requiring AI companies to test AI models and share security results with the government before they are made public.

This changed after Anthropic acted on its own initiative. Claude presented the Mythos Preview model to senior White House officialsIt was a move that exposed vulnerabilities in its software and raised concerns about the potential need for security testing of AI models before they are widely released to the public.

The White House sought to downplay the executive order as a regulatory move, emphasizing in a post Tuesday that the federal government would not conduct sweeping oversight and that the process outlined in the executive order would be voluntary.

The White House Office of Science and Technology Policy posted on

Trump’s signing of the order led to calls for Congress to take steps beyond Trump’s plan from supporters of stricter AI regulations. So far, Congress hasn’t passed any major legislation regulating artificial intelligence.

Riki Parikh, policy director of the Alliance for Secure AI, a nonprofit organization that supports safeguards for AI, said in

Progressives, including Gov. Gavin Newsom and Vermont Sen. Bernie Sanders, said the executive order was too weak and criticized Trump for backstabbing on regulation.

Some experts have suggested that the distinction between voluntary and mandatory sharing of cutting-edge technologies may be crucial.

“No company is required to formally participate, but if a developer wants to sell fringe AI systems to the federal government, participation could soon become the price of entry,” said Jessica Tillipman, a professor who studies contract law at George Washington University. Wrote an article on X.

The administration’s approach was welcomed by industry leaders, including Microsoft President Brad Smith, who said the order was “an important step toward advancing innovation while protecting the safety of the American people.”

Anthropic endorsed the order, calling it “an important step in strengthening America’s leadership in artificial intelligence.” The company said it looks forward to supporting the implementation of the programme.

Ceballos and McDaniel reported from Washington, while Christopher reported from Los Angeles. Times writer Michael Wilner contributed to this report.