The United States Cybersecurity and Infrastructure Security Agency (CISA) is utilizing Anthropic’s AI model Mythos to conduct security audits of government software repositories, according to three sources familiar with the initiative. Anthropic has a stormy relationship with the federal government, including a contentious lawsuit with the Pentagon.
Reuters reports that CISA has begun using Anthropic’s AI model Mythos to scan government code repositories for security vulnerabilities that could potentially be exploited by foreign intelligence services and cybercriminals, sources revealed on Monday. This development represents another indication of government interest in adopting the AI startup’s technology despite ongoing tensions between Anthropic and the White House.
According to one source, the scanning operations are being conducted by CISA’s Attack Surface Evaluation team, a specialized group within the agency responsible for performing digital security assessments and conducting hacking exercises across government systems. The initiative aims to identify bugs and weaknesses in government software that could create security risks.
Two sources indicated that the audits have already identified a substantial number of vulnerabilities, though they did not provide specific details about the findings. The exact scope of the code review, including how much government software has been examined and the specific nature or severity of the discovered bugs, could not be determined.
Anthropic declined to respond to inquiries about the initiative. A CISA representative stated last month that they would investigate whether information could be shared about the matter but subsequently did not respond to additional email requests.
The relationship between Anthropic and the United States government has been marked by significant conflict. The San Francisco-based company, which has confidentially filed for an initial public offering in the United States, has experienced a turbulent relationship with federal authorities. The conflict became especially serious when Anthropic filed a lawsuit against the Trump Administration over the Pentagon blacklisting the company as a supply chain risk:
The company’s complaint states that federal government contracts are already being terminated, while both current and prospective private sector agreements face uncertainty. According to the filing, these actions jeopardize hundreds of millions of dollars in near-term revenue. Anthropic CFO Krishna Rao provided additional context in a related filing, suggesting the financial impact could be far more substantial. “Across Anthropic’s entire business, and adjusting for how likely any given customer is to take a maximal reading, the government’s actions could reduce Anthropic’s 2026 revenue by multiple billions of dollars,” Rao said.
The supply chain risk designation, which was officially confirmed last week, represents an extraordinary measure historically reserved for foreign adversaries. Under this classification, defense vendors and contractors must certify they do not utilize Anthropic’s AI models, known as Claude, in any work performed for the Pentagon. This requirement effectively blocks the company from participating in defense-related contracts and projects.
More recently, the Trump Administration imposed an export ban on its Fable and Mythos AI models, which was lifted after the company sent a negotiation team to the White House:
The Trump administration implemented a ban on Friday that prohibited foreign governments, companies, and individuals from accessing Anthropic’s Fable 5 and Mythos 5 models. In response to the restrictions, Anthropic shut down access to the models entirely to ensure compliance with the new regulations. The company stated that the ban affected numerous foreign governments, companies, individuals, and even some of its own foreign-born employees.
High-level discussions took place on Saturday between administration officials and Anthropic leadership regarding Fable 5, a streamlined version of the company’s powerful Mythos model designed for public use. According to sources familiar with the talks, participants included Commerce Secretary Howard Lutnick, National Cyber Director Sean Cairncross, Anthropic co-founder and chief compute officer Tom Brown, and Sarah Heck, the company’s head of public policy.
Breitbart News social media director Wynton Hall describes AI as having great potential on the positive side, but also having the same danger as a landmine for America’s future. An AI review of government code is exactly the sort of beneficial application that Hall supports, but even such positive applications of AI may present hidden landmines. Hall has written his instant bestseller Code Red: The Left, the Right, China, and the Race to Control AI to serve as the definitive guide on how the MAGA movement can create positions on AI that benefit humanity without handing control of our nation to the leftists of Silicon Valley or allowing the Chinese to take over the world.
Read more at Reuters here.
Lucas Nolan is a reporter for Breitbart News covering issues of AI, free speech, and online censorship.