President Donald Trump signed an executive order on June 2, directing leading AI developers to voluntarily submit their most advanced models for government cybersecurity testing before public release. The order, signed in private, allows federal agencies—including the Departments of Treasury, Defense, Commerce, and Homeland Security—up to 30 days to evaluate models for potential cybersecurity risks. Participation by AI companies is voluntary, and the order explicitly prohibits mandatory licensing or preclearance requirements for AI models.
Key Details of the Order
The executive order establishes a framework for assessing AI models' "advanced cyber capabilities" and designating them as "covered frontier models" if they pose significant risks. The Treasury Department, National Security Agency, and other agencies are tasked with developing a classified benchmarking process to evaluate these models. The order also directs the creation of an AI cybersecurity clearinghouse to coordinate vulnerability scanning and patching efforts.
Background and Context
The order follows weeks of internal debate within the Trump administration. An earlier draft, which proposed a 90-day review period, was postponed after lobbying by tech industry executives and former White House AI advisor David Sacks. The final version narrows the review period to 30 days, addressing concerns about slowing innovation. The order reflects a shift in Trump's stance on AI regulation, as he had previously advocated for a hands-off approach to the tech sector.
Industry and Government Reactions
White House spokesperson Liz Huston described the order as a "common-sense approach" to balancing innovation and security. Some AI experts, such as Daniel Remler of the Center for a New American Security, noted that the order formalizes existing practices between the government and AI companies. However, the voluntary nature of the order has raised questions about its effectiveness in ensuring comprehensive cybersecurity reviews.
Long-Term Implications
The order aims to strengthen U.S. cybersecurity defenses while maintaining America's leadership in AI development. It also seeks to protect critical infrastructure from potential AI-driven threats. The voluntary framework may influence future regulatory discussions on AI governance, both domestically and internationally.