Trump Cancels Cybersecurity-Focused AI Executive Order Hours Before Signing

WASHINGTON, May 21, 2026 – A planned executive order on artificial intelligence was postponed Thursday hours before a scheduled White House signing ceremony.

Broadband Breakfast on June 3, 2026 – AI and CybersecurityHow should broadband providers be thinking about cybersecurity in an AI-driven world?Broadband BreakfastBroadband Breakfast

The president said parts of the proposal could slow innovation and undermine the United States' competitive advantage in the global AI race, citing concerns that additional oversight could become a "blocker" at a time when AI is generating investment and jobs. "I didn't like certain aspects of it, I postponed it," President Donald Trump told reporters Thursday morning at the White House.

Politico reported that the draft order would have established a voluntary federal review framework for advanced "frontier" AI models before public release. Under the proposal, federal agencies would have developed procedures for evaluating the most capable AI systems and identifying potential cybersecurity risks.

Such a pre-release review provision would have echoed a key element of the Biden administration’s 2023 AI executive order, which President Trump repealed on his first day in office in January 2025. Biden’s order invoked the Defense Production Act to compel companies to share safety test results with the government before public release if their AI systems met a certain capability threshold.

The New York Times reported that AI developers would voluntarily provide covered models to the government between 14 and 90 days before public release. Federal agencies would use that period to evaluate potential vulnerabilities and assess risks to critical infrastructure and national security systems.

The draft would apparently have directed the Office of the National Cyber Director, the White House office responsible for coordinating federal cybersecurity policy, and other agencies to develop an evaluation process within 60 days. Officials envisioned using the program to identify vulnerabilities that could affect banks, utilities, and other critical sectors before increasingly capable AI systems become widely available.

The proposal apparently also included cybersecurity initiatives across the government. The Defense Department would have been directed to secure key telecommunications and information systems, while agencies would have expanded AI deployment throughout government operations and critical infrastructure sectors, including hospitals, banks, and utilities.

The Treasury would have been tasked with establishing a voluntary partnership between AI developers and critical infrastructure operators to identify cybersecurity vulnerabilities, with support from the Cybersecurity and Infrastructure Security Agency, the National Security Agency, and the Office of the National Cyber Director.

The order emerged after growing concern within the administration that rapidly advancing AI systems could create new cybersecurity threats. Those concerns intensified following the release of Anthropic's cyber-focused Mythos model, which the company said could identify software vulnerabilities and potentially transform cybersecurity.

But Trump suggested the cure could be worse than the disease. "We're leading China, we're leading everybody," he said, "and I don't want to do anything that's going to get in the way of that lead."