Trump Signs Executive Order Creating Voluntary AI Model Review Framework
President Donald Trump has signed an executive order establishing a voluntary framework for AI companies to share frontier models with federal agencies prior to public release.
President Donald Trump signed an executive order on Tuesday establishing a voluntary framework for artificial intelligence companies to share their frontier models with the federal government before public release. According to The Verge, the directive aims to promote secure innovation and strengthen the cybersecurity of critical infrastructure. The order asserts that the domestic AI industry has thrived because the government has refused to stifle innovation with overly burdensome regulation, yet it acknowledges that emerging AI capabilities introduce new security risks that necessitate a coordinated federal response.
The order directs various federal agencies to create a framework to assess the advanced cyber capabilities of models before they reach the public. While the directive encourages cooperation, it leaves the decision to share information entirely up to AI companies. Firms that choose to participate in the voluntary program may receive certain confidentiality protections. The order explicitly clarifies that these measures should not be interpreted as a form of mandatory licensing or preclearance, distinguishing it from previous policy proposals that the administration feared might hinder competition with China.
Under the new guidelines, companies are asked to share their models with the government up to 30 days before their public release. This shift follows a period where the Trump administration had previously downplayed safety concerns and maintained a hands-off approach under former White House AI czar David Sacks. The Verge reports that 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 already committed to sharing their models with CAISI in 2024 under the previous administration.
The administration's pivot toward oversight may be influenced by the limited April rollout of Anthropic’s Mythos model. According to The Verge, the company reported that the model flagged thousands of high-severity vulnerabilities across major operating systems and web browsers. This development appears to have eased tensions between the administration and Anthropic, which had previously engaged in a legal dispute with the Pentagon regarding the use of AI for autonomous lethal weapons and mass surveillance. The order also mandates that the federal government prepare cyber defenses specifically tailored for AI, particularly concerning critical infrastructure.
The executive order has received support from groups that have historically opposed state-level AI restrictions. Americans for Responsible Innovation President Brad Carson stated that the order demonstrates the administration is taking AI vulnerabilities seriously. Similarly, Alliance for Secure AI CEO Brendan Steinhauser expressed approval of the administration’s focus on model risks. Both organizations have urged Congress to codify mandatory protections, signaling a broader industry interest in formalizing these safety standards despite the current voluntary nature of the White House framework.