The U.S. Cybersecurity and Infrastructure Security Agency (CISA) is using Mythos, the AI model designed by Claude chatbot creator Anthropic, to find security flaws in government software. The move highlights the U.S. government’s interest in using AI to strengthen cybersecurity, even though it still has a strained relationship with Anthropic.
4th of July Sale - 70% Off
- Unlock powerful investing tools and data-driven insights with TipRanks Premium for more confident investment decisions.
- Discover top stock picks and new investment opportunities through TipRanks' Smart Investor Newsletter.
CISA Uncovers Multiple Bugs After Using Mythos
According to Reuters, citing people familiar with the matter, the CISA’s Attack Surface Evaluation team is using Mythos to scan government code repositories for software bugs. These flaws could expose federal systems to cybercriminals or foreign intelligence agencies.
The audit has already uncovered many software vulnerabilities, although officials have not disclosed how many bugs were found or which agencies were affected. At the same time, Anthropic has not publicly commented on Mythos’s deployment, while the CISA has also failed to provide further details.
A CISA spokesperson also said the agency was reviewing whether it could share information about the program, but provided no further updates.
Anthropic Gains Government Support Despite Shaky Relationship
The CISA’s adoption of Mythos comes at a pivotal time for Anthropic. The AI firm has reportedly filed for a U.S. initial public offering (IPO) while managing a complex relationship with the White House.
In March, the Department of Defense (DoD) labeled Anthropic a supply-chain risk and banned the use of its model in government systems. Anthropic responded by filing two lawsuits against the Pentagon, resulting in a federal judge blocking the designation.
Later on June 9, the U.S. government also imposed an export ban on Anthropic’s Mythos over concerns that the model’s advanced bug-finding skills could be used to support cyberattacks. However, the restrictions were lifted on June 30 after Anthropic introduced extra safeguards.
Despite those tensions, demand for the company’s technology continues to grow. In fact, the National Security Agency (NSA) has been using Mythos since April of this year.
Is Anthropic Going to Be Public?
Anthropic has filed for a U.S. IPO and is widely anticipated to go public between 2026 and 2027, alongside competitors such as OpenAI. For early updates and analysts’ insights on Anthropic ahead of its public listing, visit the TipRanks Private Company Center.
