Trump’s New AI Oversight Order Seeks Security Without Stifling Innovation

President Donald Trump has signed a new executive order that increases federal oversight of the most advanced artificial intelligence systems while stopping short of imposing the kind of sweeping regulations that many in the AI industry feared. The move reflects a growing recognition that AI is becoming powerful enough to create serious national security and cybersecurity risks, while also acknowledging that excessive regulation could weaken America’s position in the global technology race.

The June 2 executive order, titled Promoting Advanced Artificial Intelligence Innovation and Security, creates a voluntary framework under which developers of the most powerful AI systems can provide the federal government access to their models for review up to 30 days before public release. The administration argues that the policy strikes a balance between security and innovation.

Trump emphasized that his administration remains committed to avoiding excessive regulation. The order states that the United States leads the world in AI because it has refused “to stifle this innovation with overly burdensome regulation” and instead has encouraged technological growth and responsible adoption.

The centerpiece of the order is the creation of a process for identifying what the administration calls “covered frontier models,” meaning advanced AI systems with potentially significant cyber capabilities.

Under the framework, companies can voluntarily work with the government to determine whether their systems qualify for review. If they do, developers may provide access to those models for up to 30 days before broader release. That review period is significantly shorter than an earlier proposal that would have allowed up to 90 days of government examination.

The order also directs multiple federal agencies to strengthen cybersecurity defenses. Among its provisions:

• Creation of an AI cybersecurity clearinghouse led by the Treasury Department.
• Expanded use of AI-powered defensive cybersecurity tools.
• Increased support for critical infrastructure such as hospitals, banks, and utilities.
• Development of classified benchmarks to evaluate the cyber capabilities of advanced AI systems.
• Greater enforcement against criminals who use AI to commit cyber crimes.

Importantly, the order explicitly states that it does not authorize a mandatory licensing system, preclearance process, or government permit requirement for new AI models. That language was included to reassure technology companies that the administration is not attempting to place federal gatekeepers between developers and innovation.

The executive order represents a notable shift from Trump’s earlier hands-off approach toward AI.

Just two weeks earlier, Trump had shelved a previous version of the order after concerns that it might slow American innovation and give China an advantage. At the time, he 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.”

According to reports, a key compromise emerged during discussions involving Treasury Secretary Scott Bessent, Defense Secretary Pete Hegseth, and AI adviser David Sacks. The review window was shortened from 90 days to 30 days, making the proposal more acceptable to those worried about regulatory creep.

The administration’s concern intensified following the release of Anthropic’s powerful AI model known as Mythos. Government officials became increasingly worried about reports that advanced AI systems could identify software vulnerabilities at a scale that human researchers cannot match. Treasury Secretary Scott Bessent warned that such models could potentially “wreak havoc” on the financial system if misused.

Who Has Been Pushing for More Oversight?

Several national security officials have advocated for stronger safeguards.

National Cyber Director Sean Cairncross was among the leading voices pushing for greater oversight of advanced AI capabilities. Treasury Secretary Scott Bessent also emerged as a major supporter after concerns grew regarding the cybersecurity implications of frontier AI models.

Outside the administration, pressure has also come from prominent conservative activists. According to reports, MAGA figures including Stephen K. Bannon, Amy Kremer, and dozens of pastors signed a letter urging Trump to adopt stronger AI vetting measures. They argued that technology companies could not be trusted “to police themselves.”

Brendan Steinhauser, chief executive of the Alliance for Secure AI Action, welcomed the administration’s move but argued that it should go further. “We would rather see this be mandatory than voluntary,” he said.

Many major technology companies have welcomed the order.

Microsoft President Brad Smith praised the effort. OpenAI’s Chief Global Affairs Officer Chris Lehane called it “an important step” and said that effective safety frameworks should be developed through democratic institutions with broad stakeholder input. Google executives also expressed support.

Michael Kratsios, director of the White House Office of Science and Technology Policy, said the order “keeps America leading in AI while putting these frontier capabilities to work strengthening our cyber defenses.”

Supporters argue that the policy acknowledges legitimate security concerns without creating a heavy-handed regulatory structure that could slow innovation.

Some critics argue that the order does not go far enough because participation remains voluntary. Others argue that it goes too far and could become the foundation for broader regulation in the future.

Former Trump AI adviser Dean Ball warned that the framework could eventually evolve into a licensing system that hinders innovation. He also questioned how the government would determine which AI systems qualify for review. Juan Londoño of the Cato Institute similarly expressed concern that giving federal officials broad discretion could create a “dangerous precedent.”

Still, even some skeptics acknowledged that the administration is attempting to address legitimate security risks while avoiding direct government control of AI development.

A Trump-Style Compromise

The new executive order ultimately reflects Trump’s effort to walk a narrow line. On one side are national security officials warning that increasingly powerful AI systems could expose dangerous vulnerabilities. On the other are technology leaders who fear regulation could undermine America’s lead over China.

By shortening the review period from 90 days to 30 days, making participation voluntary, and explicitly banning any licensing requirement, Trump appears to have crafted a compromise that increases government visibility into frontier AI development without placing major new burdens on the industry. The message from the White House is clear: keep America at the forefront of AI innovation, but keep a close eye on the risks before they become national security threats.