Trump Signs Downsized AI Cybersecurity Order

WASHINGTON, June 2, 2026 - President Donald Trump on Tuesday signed a scaled-back cybersecurity executive order narrowing federal oversight of artificial intelligence cybersecurity threats. Trump had rejected signing a broader version of the same order last month.

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

The order asks AI developers to voluntarily submit their most powerful models to a government review up to 30 days before release. The prior draft order, available from Politico – pulled hours before a scheduled signing ceremony on May 21 – had set that window at 90 days.

See Breakfast Club Membership Options!

                        See Breakfast Club Membership Options
                    

Tuesday’s executive order largely preserves the prior framework, including provisions for identifying the most powerful AI models and creating a Treasury Department-led cybersecurity clearinghouse.

The signing followed a White House meeting on Monday involving Treasury SecretaryScott Bessent, Defense Secretary Peter Hegseth and former White House Crypto and AI czar David Sacks , according to the New York Times. Sacks had opposed the draft version but backed the revised text after the review period was shortened.

Trump pulled the earlier version after expressing concern it would slow American AI development. "We're leading China, we're leading everybody," he told reporters on May 21, "and I don't want to do anything that's going to get in the way of that lead."

Under the order, the National Security Agency, the Cybersecurity and Infrastructure Security Agency (CISA) – the federal agency responsible for civilian cybersecurity – and Treasury must build a classified system for evaluating which AI models are powerful enough to warrant government scrutiny, a designation the order calls a "covered frontier model."

Once a model clears that threshold, its developer may voluntarily give the government access for up to 30 days before any public release, and work with officials to identify trusted partners for early distribution.

The order explicitly bars the government from requiring approval of AI models before they are shipped, and requires CISA to issue directives expanding AI-powered cybersecurity tools across civilian agencies.

The order also requires Treasury to establish a clearinghouse through which AI developers and critical infrastructure operators may voluntarily share information about software vulnerabilities and coordinate responses to emerging threats, and the Defense Department to

Under the order, the Committee on National Security Systems, an interagency working group, and the Defense Department are required to prioritize the cyber defense of national security and Defense information systems within 30 days, taking what the order describes as “appropriate and expeditious action” to harden them against external threats.

The attorney general must prioritize enforcement of existing federal criminal statutes, including laws covering identity fraud, unauthorized computer access and wire fraud, against anyone who uses AI to illegally access or damage computer systems, deploy AI agents to unlawfully extract data, or use AI to facilitate any other criminal activity, the order requires.

Anthropic’s Mytheos heightened concern

Concern inside the administration about AI-enabled cybersecurity threats sharpened after the release of Anthropic's Mythos model, which the AI giant said could identify software vulnerabilities and potentially transform cybersecurity defenses.

Tuesday's signing caps more than a year of executive action on AI that has consistently prioritized speed over oversight. Trump repealed**** Biden's 2023 AI order on his first day in office; that order had required developers to share safety testing with the government before release. Trump replaced it three days later with an order focused on sustaining American AI dominance.

A July 2025 AI action plan called for stripping regulations the administration said held the industry back, and a December 2025 order directed the Justice Department to challenge state AI laws deemed harmful to innovation while threatening to cut federal broadband deployment funding to states that kept such laws on the books.