WASHINGTON — President Donald Trump on Thursday postponed the signing of a long-anticipated executive order on artificial intelligence after expressing concerns about elements of the draft, saying he “didn’t like certain aspects” of the measure.
The White House had planned to unveil the order in a formal Oval Office ceremony, but the event was abruptly called off hours before it was scheduled to take place. Trump later told reporters that he had put the signing on hold because he wanted changes to the text and was wary of any policy that could undercut the United States’ lead in the fast-moving global AI race.
“I didn’t like certain aspects of it,” Trump said, according to multiple reports. “I postponed it.”
The decision highlights how central artificial intelligence has become to the administration’s technology agenda, while also underscoring the political tension between promoting innovation and imposing guardrails on a sector seen as both economically transformative and potentially disruptive. The draft order reportedly dealt with issues ranging from security and oversight to the handling of AI-generated content, including guidance tied to watermarking manipulated images and videos.
Order aimed to shape federal AI oversight
Although the full text of the order was not immediately released, people familiar with the matter said it was expected to give the federal government more authority to review certain advanced AI models before public release. Such a framework would have marked a significant step toward direct federal involvement in an industry that has largely operated under a patchwork of voluntary commitments, agency guidance and state-level rules.
That kind of approach has drawn support from some lawmakers and regulators who argue that powerful AI systems should be tested more thoroughly before they are widely deployed. Others in the tech industry have warned that broad government review requirements could slow development, raise compliance costs and discourage smaller companies from competing with larger rivals.
Trump’s remarks suggested that competition with China remained a major factor in his thinking. He has repeatedly framed artificial intelligence as a strategic domain in which the United States must move quickly to maintain leadership. By postponing the signing, Trump signaled that any federal AI policy would need to balance national security concerns with his broader push to keep American companies ahead of foreign competitors.
Balancing innovation and safeguards
The administration’s pause came amid a broader national debate over how aggressively Washington should regulate artificial intelligence. Supporters of stronger safeguards say the technology can be used to spread misinformation, create realistic deepfakes, automate cyberattacks and displace workers across multiple industries. They argue that without some level of oversight, the risks of powerful AI systems could outpace the benefits.
At the same time, many technology executives and venture investors have cautioned against rules that they believe would freeze innovation or push development overseas. They say the U.S. needs a light-touch regulatory approach to stay competitive in a sector where rapid iteration is often essential.
The timing of Trump’s postponement also came as state governments and federal agencies continued to examine the impact of AI on employment, privacy and public safety. California Governor Gavin Newsom, for example, has been reported to be preparing state agencies for possible workforce disruptions related to AI, reflecting growing concern that automation could reshape labor markets faster than policymakers can respond.
That parallel effort illustrates the patchwork nature of AI governance in the United States. While the federal government considers nationwide standards, states are moving ahead with their own frameworks on issues such as data privacy, deepfakes, election integrity and worker protections.
White House has faced pressure on AI policy
Trump’s delay is likely to intensify scrutiny of the White House’s AI agenda. Any executive order from the president would be closely watched by both Silicon Valley and Capitol Hill, especially if it includes requirements for testing, disclosure or content labeling. A revised order could still emerge after further review, but the episode shows that the administration is not prepared to move quickly on a policy it views as technically and politically delicate.
The draft also appears to have touched on cybersecurity concerns, another major area of debate as AI systems become more integrated into business and government operations. Experts have warned that advanced models can be used to generate convincing phishing content, automate vulnerability discovery and assist malicious actors in scaling attacks. At the same time, defenders argue AI can improve threat detection and response if deployed responsibly.
In recent months, U.S. officials have increasingly treated AI as both an economic opportunity and a national security issue. The technology is expected to transform sectors including health care, finance, defense, manufacturing and education. But concerns about misinformation, intellectual property, bias, and labor displacement have made it difficult for policymakers to settle on a single regulatory path.
Trump’s decision to stop the signing ceremony does not necessarily mean the administration has stepped back from AI oversight altogether. Rather, it suggests the White House is still negotiating the terms of its approach, with the president himself taking an active role in determining which provisions should remain in the final version.
What comes next
For now, the postponed order leaves open the question of when — and in what form — the administration will act. If the White House revises the draft and moves forward, the final order could provide an early blueprint for how Trump intends to manage one of the defining technologies of the decade.
Industry executives, civil liberties advocates and lawmakers are expected to watch closely for any changes. A softer version could emphasize voluntary standards and federal coordination, while a more aggressive revision could expand government review and content-safeguarding requirements.
Either way, the episode makes clear that artificial intelligence is no longer a niche policy issue. It has become a central test of how the federal government will protect consumers, preserve national security and foster innovation at the same time.
For Trump, the decision to delay the order suggests he wants a version that he believes better aligns with his priorities: maintaining American dominance in AI, avoiding unnecessary bureaucratic barriers and ensuring that federal policy does not slow the country’s technological momentum.