WASHINGTON D.C.: The U.S. Cybersecurity and Infrastructure Security Agency (CISA) is using Anthropic's artificial intelligence model Mythos to scan government software for security flaws, according to three people familiar with the matter, highlighting growing government adoption of the company's AI tools.

The agency is using Mythos to examine government code repositories for vulnerabilities that could be exploited by foreign intelligence agencies or cybercriminals, the sources said.

Anthropic did not respond to questions about the initiative. A CISA spokesperson said last month that the agency would check whether there was information to share but did not respond to subsequent requests for comment.

According to one of the sources, the work is being carried out by CISA's Attack Surface Evaluation team, which conducts digital security assessments and hacking exercises across the U.S. government.

Two of the sources said the audits have already uncovered a large number of vulnerabilities, although they did not provide details. It could not be determined how much government code had been reviewed or the nature and severity of the flaws identified.

The project comes as Anthropic continues to navigate a complicated relationship with the U.S. government.

The company, which has confidentially filed for a U.S. initial public offering, came under pressure earlier this year after refusing to remove safeguards that prevented its AI systems from being used for autonomous weapons or domestic surveillance. The dispute led the Pentagon in February to designate Anthropic as a supply-chain risk, a classification previously used for foreign companies suspected of facilitating espionage.

A judge blocked the designation in March, and tensions have eased since Anthropic privately released Mythos, an AI model designed to identify and exploit cybersecurity vulnerabilities.

The National Security Agency has also been using Mythos since at least April despite the earlier blacklist, according to Axios. The New York Times reported last month that NSA analysts had been testing the model in classified environments and were impressed by its capabilities.

When Anthropic later released a public version of Mythos, called Fable, with cybersecurity safeguards, the White House ordered the company to block access for foreign users. That decision led to a temporary global shutdown of the model before restrictions were eased last week.

The NSA and the White House did not immediately respond to requests for comment.