Trump signs voluntary AI model review order following industry pushback
The directive aims to bolster cybersecurity for critical infrastructure while explicitly ruling out mandatory licensing, aligning with recent commitments from major tech firms to the Commerce Department.
President Donald Trump signed an executive order on Tuesday establishing a voluntary framework for artificial intelligence companies to share frontier models with the federal government prior to public release. The directive, developed by the Office of the National Cyber Director, requires developers to submit their most powerful models for government review up to 30 days before launch. This timeframe represents a significant reduction from an initial proposal of 14 to 90 days, following industry pushback and objections from former White House AI czar David Sacks.
The order aims to promote secure innovation and strengthen cybersecurity for critical infrastructure, allowing companies discretion over sharing information while offering confidentiality protections. It explicitly states that the framework should not be interpreted as mandatory licensing or preclearance. The directive acknowledges that while the US AI industry has succeeded by avoiding overly burdensome regulation, new capabilities introduce security risks that require assessment.
Major technology firms have already begun aligning with these expectations. Google, Microsoft, and xAI agreed last month to allow pre-release review by the Commerce Department’s Center for AI Standards and Innovation (CAISI). OpenAI and Anthropic had previously agreed to share models with CAISI in 2024 under the Biden administration. The signed order directs several federal agencies to develop a framework to assess the advanced cyber capabilities of AI models.
The shift in policy follows a last-minute postponement of a previously planned order that Trump feared could hinder competition with China. Advocacy groups have praised the new directive, with Americans for Responsible Innovation President Brad Carson noting the administration is taking AI vulnerabilities seriously. Alliance for Secure AI CEO Brendan Steinhauser also expressed pleasure at the administration’s focus on model risks, though both leaders urged Congress to codify mandatory protections.
One factor influencing this regulatory posture may be Anthropic’s April rollout of its Mythos model, which flagged thousands of high-severity vulnerabilities in major operating systems and web browsers. The order reflects a willingness to employ oversight of AI companies, even as the administration maintains a generally hands-off approach under former AI czar David Sacks. The executive order marks a notable step in balancing innovation with the need for robust cyber defenses.
