President Donald Trump signed a landmark executive order Tuesday requiring AI companies to give the government early access to their most powerful models for review.
Subscribe to read this story without ads
Get unlimited access to ad-free articles and exclusive content.
The order, signed in private, directs federal agencies, including the Department of Defense, the Treasury Department, and the Cybersecurity and Infrastructure Security Agency, to strengthen cybersecurity defenses of the nation’s critical infrastructure and lays the groundwork for the federal government to test and vet the most powerful AI systems for safety issues before deploying them.
The test will rely on voluntary cooperation from major US AI companies, including Anthropic, OpenAI, and Google. The order explicitly prohibits the government from imposing mandatory licensing or pre-approval requirements for new AI models, making it a request rather than a rule.
“Advanced AI capabilities strengthen the nation, but they also introduce new national security considerations and require coordinated action across the executive branch and agencies,” the executive order states. The order states that the Trump administration will “work closely with industry to ensure that the best and safest technology is rapidly deployed to combat all threats to our nation.”
The executive order was originally scheduled to be signed in late May, two people familiar with the matter said. The White House invited technology company CEOs to the signing ceremony and held a press conference with senior officials on the morning of the planned announcement.
But President Trump withdrew the order at the last minute, later telling reporters that the order could hinder U.S. companies’ ability to compete with Chinese companies.
“We lead China, we lead all countries, and we don’t want to do anything that gets in the way of that lead,” Trump said at the time.
Tuesday’s order stipulates that the voluntary testing mechanism will provide state-of-the-art AI models to the federal government “up to 30 days in advance of any plans to release such models to other trusted partners.”
The executive order has been in the works for months. In April, AI company Anthropic’s new Mythos Preview model shocked Washington with its superhuman ability to discover critical vulnerabilities in the world’s most used operating system.
Government officials, including Treasury Secretary Scott Bessent and Vice President J.D. Vance, are increasingly focused on managing the spread of AI and limiting the threat to U.S. national security and economy.
