ON June 2, 2026, President Donald Trump signed an executive order inviting developers of advanced AI models to voluntarily submit their models for cybersecurity review by the US government prior to release. This order establishes a framework that allows access to these 'covered frontier models' for up to 30 days without creating mandatory licensing requirements.
Concerns about the security implications of frontier AI models, specifically those capable of exploiting software vulnerabilities, have driven this initiative. The order also outlines measures for strengthening national security infrastructures and creating an 'AI cybersecurity clearinghouse' to manage vulnerabilities. While the industry response is cautious, highlighting potential accountability issues, there is support for the concept of voluntary security programs.