Trump signs AI safety order, tests models before release

voluntary review – President Trump signed an executive order directing federal agencies to set benchmarks for AI models’ cyber capabilities and to build an “AI cybersecurity clearinghouse,” while asking top AI companies to voluntarily submit their most powerful models for govern

On Tuesday, President Trump signed a long-awaited executive order that puts the government closer to the most advanced artificial intelligence than his administration had signaled it would be just weeks ago.

The order asks leading AI companies to voluntarily submit their most powerful models for government testing up to 30 days before they’re released to the public. The White House describes the shift as necessary to address the security threats that advanced AI capabilities can introduce.

“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 adds that as the capabilities evolve. “my Administration will continue to work closely with industry to ensure that the best and most secure technology is deployed rapidly to confront any and all threats to our country.”.

Alongside the voluntary testing request, the executive order directs federal agencies to develop benchmarks to assess AI models’ cyber capabilities. It also calls for the creation of an “an AI cybersecurity clearinghouse” that would review and share information on vulnerabilities. Federal officials are tasked with strengthening government security defenses as these systems move from labs toward everyday use.

The timing makes the decision feel more urgent than the paperwork itself. The order was expected last month. but the White House scrapped plans to sign it over concerns that it would interfere with AI innovation. Trump said at the time he worried the order would stifle American companies’ lead in the global race amid competitive pressure from China.

In that earlier draft, the government would have had up to 90 days to review advanced models before release. In the final order, that review timeline is cut to 30 days.

The new executive order also lands in a larger struggle inside the administration over how far the federal government should go to regulate a fast-moving technology. The Biden White House pushed for federal oversight. while Trump has generally sought to minimize regulation— including at the state level— even as safety concerns have proliferated.

But recently. the development of more powerful AI models has spooked some federal officials. pushing the White House to reverse course and back safety measures. One trigger was Anthropic’s April announcement that it was limiting the release of its new Mythos Preview model because of its ability to identify and exploit software security vulnerabilities. That disclosure reverberated across Silicon Valley and Washington.

The executive order, however, is careful about how far it goes. It relies on voluntary cooperation from the tech companies leading AI development, including OpenAI, Anthropic, and Google. It also spells out that the testing framework is not a government licensing system. The order states that “nothing in this section shall be construed to authorize the creation of a mandatory governmental licensing. preclearance. or permitting requirement for the development. publication. release. or distribution of new AI models.” Any regulations around the technology would have to come from Congress.

What’s changed is not just the policy language, but the posture behind it: a White House that once leaned away from interference is now asking for access, benchmarks, and a new security information hub—while trying to keep the hand of regulation off the table.

Trump executive order AI safety voluntary model testing AI cybersecurity clearinghouse cyber capabilities benchmarks Anthropic Mythos Preview OpenAI Google executive departments Congress regulation