Accenture AI Agent Breach Exposes DOD Credentials

• •An Accenture employee connected a secure DOD system to a personal AI account, exposing credentials.
• •The AI agent ran undetected for two months, sending sensitive data to a flagged third-party site.
• •Accenture failed to report the incident within the required 72-hour timeframe.

An Accenture employee connected a secure Department of Defense (DOD) system to their personal Claude Code account, allowing Anthropic’s AI coding agent to run on the Pentagon network for two months. The agent reportedly sent sensitive information, including login credentials and a network map, to a high-risk third-party website. The affected system operates on the Non-classified Internet Protocol Router Network (NIPRNet), which holds controlled unclassified information (CUI). Security experts have flagged this incident as a significant national security threat, with concerns about potential exploitation of sensitive data. Accenture did not initially report the incident to the government, violating DOD protocols requiring reporting within 72 hours. The DOD and Accenture have both stated that investigations found the allegations to be without merit, but details about these investigations remain unclear. The breach raises serious concerns about the security of DOD systems amid ongoing geopolitical tensions.