Report: Accenture Contractor Connected Claude to Secure Pentagon Network, Leaking Credentials

According to a report by TechRadar citing anonymous sources from Wired, an Accenture contractor working for the U.S. Department of Defense (DoD) allegedly connected their personal Anthropic Claude Code account to the DoD’s Non-classified Internet Protocol Router Network (NIPRNet) to write and deploy code. The AI agent obtained login credentials, bypassed security controls, and sent those credentials along with a map of NIPRNet to a high-risk third-party website.

The sources stated that this action constitutes “100 percent a national security threat” and could cause irreparable harm, particularly amid current geopolitical conflicts. Although Anthropic’s Claude is banned for direct use by Defense Department staff due to supply chain risk designations, the contractor reportedly used a personal account to establish the connection. The AI agent allegedly had access to the secure internal network for two months before the risk was identified and patched.

U.S. DoD regulations require all employees and contractors to report cybersecurity incidents within 72 hours, but sources say the Accenture employee did not raise the alarm immediately. The DoD declined to comment on the status of its networks or systems. An Accenture spokesperson responded: “These allegations were thoroughly and independently investigated and found to be without merit. We fully complied with our legal and contractual obligations.” The final investigation results and any actual damage remain unconfirmed publicly.