Executive order sets voluntary cyber reviews for advanced AI

President Donald Trump signed an executive order on June 2, 2026, introducing voluntary cybersecurity reviews for developers of advanced artificial intelligence models. The order follows a delay from May 21, when Trump postponed signing due to concerns that mandatory reviews could hinder AI-driven job growth and U.S. competitiveness, particularly against China. The new framework allows developers to submit frontier AI models to federal agencies—including the Treasury Department, National Security Agency (NSA), and the Department of Homeland Security’s Cybersecurity and Infrastructure Security Agency (CISA)—up to 30 days before public release. Agencies will establish benchmarks to identify qualifying models and work with developers to designate trusted partners for early access, aiming to strengthen cybersecurity for critical infrastructure. The order also mandates the creation of an AI cybersecurity clearinghouse, developed in collaboration with industry, to coordinate vulnerability scanning, validation, and patch distribution. Additionally, CISA must issue Binding Operational Directives to enhance federal cyber defenses, including AI-enabled tools, while the Defense Department and Committee on National Security Systems will prioritize securing government information systems. Senator Josh Hawley (R-Mo.) praised the measure but called for mandatory reporting, citing his September 2025 legislation with Senators Richard Blumenthal (D-Conn.) and Marsha Blackburn (R-Tenn.) to require mandatory AI reviews by the Energy Department. Meanwhile, Brendan Steinhauser, CEO of the Alliance for Secure AI, criticized the voluntary approach, arguing it falls short of addressing critical cybersecurity risks. The executive order reflects ongoing tensions between fostering AI innovation and mitigating national security threats, with agencies tasked to balance rapid deployment with robust cybersecurity safeguards.