President Donald Trump signed an executive order requiring federal review of new AI models before public release, establishing a 30-day evaluation period and creating an AI cybersecurity clearinghouse to address security vulnerabilities. This move reflects increased government oversight amid security concerns, including tensions with AI companies like Anthropic, and is supported by a newly appointed AI advisory panel featuring prominent tech leaders.
On Tuesday, President Donald Trump signed a highly anticipated executive order aimed at increasing federal oversight of new artificial intelligence (AI) models before they are released to the public. This marks a notable shift from his administration’s earlier preference for a more deregulated approach to AI technology. The order mandates that companies submit their AI models for a 30-day review to evaluate their “advanced cyber capabilities” and determine if they qualify as “covered frontier models,” though the criteria for this designation remain unspecified.
The executive order outlines a process where the government can review AI models up to 90 days before their public release, with a 30-day window specifically allocated for the review itself. During this period, the government may select “trusted partners” to receive early access to these models. This approach is designed to ensure that potential risks are identified and mitigated before AI technologies become widely available, reflecting growing concerns about the security and ethical implications of advanced AI systems.
A key component of the order is the establishment of an AI cybersecurity clearinghouse, led by the Treasury Department in collaboration with the AI industry and critical infrastructure operators. This clearinghouse will focus on identifying and addressing security vulnerabilities discovered in AI models, aiming to enhance the overall safety and reliability of AI deployments. The administration emphasized its commitment to deploying secure technology rapidly to protect national interests and directed the attorney general to prioritize prosecuting AI-related crimes.
The executive order appears to be influenced by recent tensions between the Trump administration and AI companies, notably a contract dispute between Anthropic and the Pentagon. Anthropic’s Claude model, which reportedly could exploit security flaws in major operating systems and browsers, triggered heightened scrutiny and several meetings with government officials. Defense Secretary Pete Hegseth even labeled Anthropic as a supply chain risk, signaling increased caution toward AI providers perceived as potential security threats.
In addition to the executive order, Trump has taken steps to bolster AI governance by appointing over a dozen members to a new AI advisory panel earlier this year. This panel includes prominent tech leaders such as Meta’s Mark Zuckerberg, Oracle’s Larry Ellison, and Nvidia’s Jensen Huang. The group is tasked with advising the president on science, technology, education, and innovation policy, underscoring the administration’s recognition of AI’s strategic importance and the need for informed oversight.