Trump Signs Executive Order for Voluntary AI Model Review Before Public Release

President Donald Trump signed a significant executive order on Tuesday, June 2, 2026, establishing a "voluntary framework" for artificial intelligence companies. This new directive encourages firms to share their advanced AI models with the federal government for review before their public release. The primary objectives are to foster secure innovation and bolster the cybersecurity of critical national infrastructure, acknowledging the inherent security risks that accompany new AI capabilities. Under this framework, AI companies retain the discretion to decide whether to submit their models for pre-release assessment. However, those that opt to participate would be eligible for specific confidentiality protections. The order mandates several federal agencies to develop a comprehensive system for evaluating the advanced cyber capabilities of these AI models. It also emphasizes the federal government's responsibility to enhance its own cyber defenses against AI threats, particularly concerning vital infrastructure. The current order suggests a review period of up to 30 days before public release, a slight adjustment from earlier proposals that ranged from 14 to 90 days. This executive order marks a notable evolution in the administration's approach to AI governance. It follows a previous postponement by President Trump of an earlier order, which he feared might impede American competitiveness against China. While the administration had previously adopted a more "hands-off" stance, downplaying safety concerns, this new order explicitly states it is not a form of mandatory licensing or preclearance, yet clearly signals a willingness for greater oversight. Major tech giants like Google, Microsoft, and xAI have already agreed to pre-release reviews by the Commerce Department’s Center for AI Standards and Innovation (CAISI), building on similar agreements made by OpenAI and Anthropic under President Biden in 2024. A key catalyst for this shift appears to be Anthropic's limited April rollout of its powerful Mythos model. The company reported that Mythos had identified "thousands of high-severity vulnerabilities" across major operating systems and web browsers, underscoring the potential risks of advanced AI. This revelation seemingly created an opportunity to ease tensions between Anthropic and the administration, especially after past legal disputes regarding AI use in autonomous weapons and mass surveillance. This order reflects a growing recognition within the administration of the importance of balancing rapid innovation with necessary security safeguards to protect society and critical infrastructure from potential harmful uses of AI. The order has largely been met with approval from various advocacy groups. Brad Carson, President of Americans for Responsible Innovation, praised the move, stating it shows the Trump administration is "taking AI vulnerabilities seriously." Similarly, Brendan Steinhauser, CEO of Alliance for Secure AI, expressed satisfaction with the administration's acknowledgment of these risks. Both leaders have urged Congress to take further steps by codifying mandatory protections, suggesting that while this voluntary framework is a positive step, more robust legislative action may be necessary to fully address the complex challenges posed by advanced AI.