President Donald Trump last week abruptly canceled the signing of a sweeping artificial intelligence (AI) executive order that would have mandated strict government evaluation for frontier AI models before their public release. The regulation would have forced major technology firms, including Google, Meta, OpenAI, and Microsoft, among others, to submit their most powerful systems to federal agencies for safety vetting. The sudden cancellation occurred just hours before the scheduled Oval Office event, following the President's recent high-profile state visit to China.
According to a report by The New York Times, Trump altered his stance on the policy out of a desire to protect America's competitive advantage over foreign adversaries. "I think it gets in the way of—you know, 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," Trump told reporters when explaining the decision to shelve the policy.
Details of the Draft Executive Order
The draft executive order, obtained by Politico, was titled "Promoting Advanced Artificial Intelligence Innovation and Security." It emphasized the United States' leadership in AI and aimed to avoid burdensome regulation while addressing national security considerations. Key provisions included upgrading federal cybersecurity systems, creating a classified benchmarking process for AI models, and establishing a voluntary framework for AI developers to engage with the government.
Upgrading American Systems for Advanced AI
The order directed several actions within 30 to 60 days, including:
- Prioritizing cyber defense of National Security Systems by the Committee on National Security Systems.
- Directing the Secretary of War to prioritize cyber defense of Department of War information systems.
- Releasing Binding Operational Directives from the Cybersecurity and Infrastructure Security Agency (CISA) to expedite civilian federal government cyber defense, expand AI-enabled defensive tools, and facilitate access to cybersecurity services.
- Forming an AI cybersecurity clearinghouse involving the Treasury, NSA, and CISA to coordinate vulnerability scanning and patching.
- Expanding the U.S. Tech Force Information Cybersecurity Specialist hiring pathways.
Secure Frontier Model Deployment
Within 60 days, the order required the development of a classified benchmarking process to assess AI models' advanced cyber capabilities and designate certain models as "covered frontier models." It proposed a voluntary framework for developers to engage with the government, provide access to models before release, and collaborate on selecting trusted partners. Importantly, the order explicitly stated that nothing in it would authorize mandatory governmental licensing, preclearance, or permitting for AI model development or distribution.
Protection Against Criminal Actors
The Attorney General was directed to prioritize enforcement of federal laws against individuals using AI for illegal computer access or damage, including unauthorized access to information systems or using AI agents for criminal purposes.
Reasons for Cancellation
Trump's decision to cancel the order stemmed from his belief that strict regulation could hinder U.S. innovation and cede ground to China. The draft order itself reflected a balance between promoting innovation and addressing security concerns, but ultimately the President chose to avoid any potential impediment to American AI leadership. The cancellation came after his visit to China, which may have influenced his perspective on global competition.
The abrupt reversal highlights the ongoing debate over AI regulation in the U.S., where policymakers grapple with ensuring safety without stifling technological progress. Industry reactions were mixed, with some companies welcoming the lighter regulatory touch while others expressed concerns about potential risks from unvetted AI systems.
