US Agencies Deploy AI for Software Audits

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) is now utilizing Anthropic's sophisticated AI model, Mythos, to conduct security audits of government software. This initiative involves scanning code repositories for potential bugs and vulnerabilities that could be exploited by foreign adversaries or cybercriminals. The AI's deployment signifies a notable embrace of advanced artificial intelligence by U.S. government agencies for enhancing national cybersecurity measures.

This adoption occurs amidst a complex relationship between Anthropic and the U.S. government. Previously, the company faced scrutiny over its AI's safeguards. However, its Mythos model, known for identifying and exploiting cybersecurity weaknesses, has reportedly impressed agencies like the National Security Agency (NSA), which has been using it in classified settings. Audits conducted by CISA's Attack Surface Evaluation team have reportedly uncovered a significant number of vulnerabilities, though specifics remain undisclosed.