President Donald Trump signed a historic decree On Tuesday, it lays the groundwork for federal testing of the world’s most powerful AI systems before they are released to the public.
The order, signed privately, directs federal agencies – including the Pentagon, Treasury and the Cybersecurity and Infrastructure Security Agency – to strengthen the nation’s cybersecurity defenses for critical infrastructure, and sets out a mechanism for the federal government to test and vet the most powerful AI systems for security issues before they are deployed.
The tests would rely on the voluntary collaboration of major American AI companies, such as Anthropic, OpenAI and Google. The order explicitly prohibits the government from creating a compulsory licensing or prior authorization requirement for new AI models, making the government a request and not a rule.
“Advanced AI capabilities make our nation stronger, but also introduce new national security considerations that require coordinated action across executive departments and agencies. » the decree says. The order states that the Trump administration “will work closely with industry to ensure that the best and most secure technology is deployed quickly to address all threats to our country.”
The order would also direct a coalition of administration officials and offices spanning the Department of the Treasury, the National Security Agency, and the Department of Homeland Security to create an AI cybersecurity clearinghouse with collaboration from industry partners to centralize knowledge about software vulnerabilities and fixes relevant to these issues.
The decree was initially scheduled to be signed on May 21, according to two sources familiar with the matter. The administration invited tech CEOs to attend the signing ceremony and held a news conference with top officials the morning of the planned announcement. But the May signing and ceremony were canceled at the last minute. Later in the day, Trump told reporters that the order could have harmed the competitiveness of American companies compared to Chinese companies.
“We lead China, we lead everyone, and I don’t want to do anything that’s going to stand in the way of that lead,” Trump said at the time.
Tuesday’s order states that the voluntary testing mechanism would allow the federal government to access cutting-edge AI models “up to 30 days before it considers releasing those models to other trusted partners.” The original decree stated that companies would be able to provide their models to the government up to 90 days before their public release.
The order also directs the attorney general to prioritize the prosecution of crimes involving AI, with a particular focus on cybercrimes. The order also asks prosecutors to focus on people using AI agents, or autonomous AI systems, to “unlawfully access data or information that is then used for criminal or unlawful purposes.”
The decree has been in preparation for months. In April, AI company Anthropic’s new Mythos Preview model sent shockwaves through Washington with its superhuman ability to find critical and serious vulnerabilities in the world’s most widely used operating systems.
As the capabilities of AI systems have continued to increase over the past few years, experts said that Mythos Preview represented a significant advancement in cybersecurity capabilities in particular. Although the model could help companies identify weaknesses in their cybersecurity programs, experts — and many in government — worried the model could help bad actors or hackers exploit bugs in software systems.
To help businesses close gaps in their cybersecurity defenses, Anthropic launched a program called Project Glasswing to give early access to Mythos Preview to around fifty partners, including Cisco, Microsoft and Nvidia. On Tuesday, Anthropic announced it would expand access to 150 organizations in more than 15 countries.
OpenAI Anthropogenic Competitor launched a similar effort in April.
Since taking office, the Trump administration has largely championed a light-touch approach to AI regulation, pushing back against federal proposals and state laws that it says will negatively affect U.S. AI innovation.
Former White House AI czar and venture capitalist David Sacks has onerous state regulations are regularly cited as a threat to the US AI industry. According to reportsSacks advocated withdrawing the previous iteration of the executive order at the last minute.
Senior administration officials, including Treasury Secretary Scott Bessent And Vice President JD Vanceare increasingly invested in regulating the spread of AI and limiting threats to security and the American economy. “We want to make sure we protect people” from AI risks, Vance said at a recent news conference, mentioning data and Americans’ privacy as specific concerns.
Although the Trump administration aims to foster closer collaboration with leading U.S. AI companies through Tuesday’s executive order, it remains technically locked in a legal battle with Anthropic that erupted several months before the limited release of Mythos Preview.
Anthropic is adamant that its systems should not be used directly in lethal autonomous weapons or for mass national surveillance. The Pentagon balked at these stipulations, leading the Defense Department and President Trump to cut ties with the company. Although a court ruled in favor of Anthropic in California, the case remains pending in Washington, DC.
