WASHINGTON — President Donald Trump abruptly canceled an Oval Office event hours before he was scheduled to sign an executive order granting the government the power to test frontier artificial intelligence models before their public release. Trump told reporters he decided against signing the order because he "didn't like certain aspects of it."

"I really thought [the order] could have been a blocker," Trump 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."

The order would have established a framework for the government to vet the national security risks of the most advanced AI systems before their public release. It was being characterized as a voluntary collaboration with U.S.-based tech companies including Anthropic, OpenAI, and Google, and aimed for the government to identify security vulnerabilities revealed by AI models and patch problems in its systems to help protect banks, utilities, and other sensitive industries from cyberattacks. The Information reported that the government sought to evaluate AI models up to 90 days prior to release, while AI labs pushed for a much shorter timeline of 14 days.

Trump had hoped top executives from leading AI firms would attend the signing. Some CEOs had been given only 24 hours' notice and could not attend, while other executives who rearranged their schedules were midair on their way to the Oval Office when they found out the trip was canceled. OpenAI supported the signing. The tech industry lobbied against the order, fearing that safety testing could delay model launches or require changes that would set back model development.

Elon Musk, Mark Zuckerberg, and David Sacks urged Trump to call off the signing. Musk denied on X that he had a hand in the cancellation, writing "this is false" and stating that he did not know what was in the executive order. The Information reported that Sacks's special government employee designation expired in March, though he continues to visit the White House weekly.

Members of Trump's administration had recommended safety testing after Anthropic flagged cybersecurity risks with its latest model, Mythos. In April, Federal Reserve Chair Jerome Powell and Treasury Secretary Scott Bessent convened an urgent meeting with Wall Street CEOs to warn about cybersecurity risks posed by Claude Mythos.

Lizzi C. Lee, a fellow at the Asia Society Policy Institute's Center for China Analysis, said Trump appears to be navigating the same AI safety dilemma as China: how to guard against national security risks without slowing the development of frontier models. Lee said the impact of Trump's order depends on how heavy the review process becomes, and that if safety testing focuses narrowly on national security, it probably would not slow leading U.S. labs much. She also said that parallel to the AI race is a separate race to figure out who can govern powerful AI without choking off innovation.

Since retaking office, Trump has taken a hands-off approach to regulating AI and had pledged to undo the AI safety regulations set by President Joe Biden. His Commerce Department signed agreements with Google, Microsoft, and Musk's xAI to evaluate their most powerful AI models before their public release, though the announcement later disappeared from the department's website. In February, Trump ordered all U.S. agencies to stop using Anthropic's chatbot Claude after a public clash between the Pentagon and CEO Dario Amodei.