The Trump administration is in early discussions about whether advanced AI models, like the new Anthropic Mythos, should be vetted before they are released to the public, according to reports from the New York Times, Wall Street Journal and Axios. The discussions focus on systems that are capable of facilitating cyberattacks, particularly models that could help users identify and exploit software vulnerabilities.
See also: Anthropic and Wall Street create $1.5 billion pipeline to private capital markets
Officials are considering several options, including formal pre-release review processes and government-led testing for higher-risk systems. No proposal has been finalized and no timeline has been set.
The discussions mark a shift in tone, if not yet policy. On January 20, 2025, Donald Trump’s first day back in office during his second term, he revoked Biden’s Executive Order 14110 on the Safe and Trustworthy Development and Use of Artificial Intelligence. Three days later, he issued his own order, “Removing Obstacles to American Leadership in Artificial Intelligence,” marking a significant shift away from the Biden administration’s emphasis on oversight and risk reduction toward a framework focused on deregulation and fostering AI innovation.
Among the things the order effectively ended: The Biden framework had introduced mandatory “red teams” for high-risk AI models, enhanced cybersecurity protocols, and monitoring requirements for AI used in critical infrastructure. The new discussions suggest that some security risks — particularly those associated with offensive cyber capabilities — require a more interventionist stance, even as the administration remains generally opposed to sweeping AI regulation.
The discussion follows the recent introduction of Anthropic Mythos, a model that the company described as representing a defining moment for cybersecurity. Anthropic has said that Mythos Preview has found thousands of high-severity vulnerabilities, including some in every major operating system and web browser, and that AI models have reached a level of coding proficiency where they can outperform all but the most skilled humans at finding and exploiting software vulnerabilities.
See also: The White House opposes the expansion of the Anthropic Mythos
In one benchmark, the company reported significantly higher success rates compared to previous models.
Anthropic has not released the model publicly. Instead, it has launched Project Glasswing, committing up to $100 million in usage credits to a select group of technology and cybersecurity companies to use Mythos for defensive purposes — finding and fixing vulnerabilities before malicious actors can exploit them. Anthropic has also briefed the Cybersecurity and Infrastructure Agency, the Department of Commerce, and other stakeholders about the potential risks and benefits of the Mythos Preview.
OpenAI has developed a comparable model and released it to a small set of companies through an existing trusted access program.
Pre-release evaluation of AI models is not a new idea, but it remains poorly defined within U.S. policy. Biden’s executive order that Trump revoked required developers of larger AI systems to notify the government and share the results of security tests before deployment — one of several provisions that the Trump administration has characterized as burdensome barriers to innovation.
The institutional landscape has also changed. The U.S. AI Security Institute, created under Biden’s order to conduct pre-deployment assessments and housed within the National Institute of Standards and Technology, was significantly reorganized after Trump took office. In June 2025, the organization was renamed the Center for AI Standards and Innovation, and its mission was revised.
See also: Claude: Integration with Adobe, SketchUp and other creative apps
Commerce Secretary Howard Latnick called the change a rejection of what he called using security as a pretext for censorship and regulation. The renamed center’s mandate now includes leading unclassified assessments of AI capabilities that may pose national security risks, with a stated focus on proven risks such as cybersecurity, biosecurity and chemical weapons, potentially positioning it to play a role in any future review process.
