CISA Deploys Anthropic Mythos AI to Audit Government Code Repositories for Security Vulnerabilities

CISA is using Anthropic's Mythos AI model to audit government software, three people familiar with the matter told Reuters 1. The Attack Surface Evaluation team, a unit that runs digital security assessments across government, is conducting the scans, checking code repositories for bugs that could expose systems to foreign spies and cybercriminals, one source said 1. Two sources said the audits have already uncovered a large number of vulnerabilities, though Reuters could not establish how much code has been reviewed or the severity of the bugs found 1. The initiative comes as Anthropic's relationship with the government has been strained since February, when the Pentagon imposed a formal supply-chain risk designation on the company over its refusal to strip safeguards blocking use of its AI for autonomous weapons or domestic surveillance; a judge blocked that designation in March, and tensions have eased since the private release of Mythos 1. Anthropic did not respond to questions about the initiative, and a CISA representative who said last month he would check on the matter did not respond to further emails 1.