Bill > HR10189

This bill, known as the Defense AI Reliability and Reporting Act, directs the Secretary of Defense to establish a centralized program within the Department of Defense (DoD) to track, analyze, and fix safety incidents and security weaknesses involving artificial intelligence (AI), which refers to machine-based systems designed to make decisions or operate with varying levels of autonomy. Under this program, military personnel, civilian employees, and contractors are required to promptly report any "covered AI incidents" (events where an AI system causes unintended harm, acts outside safety guardrails, fails to respond to human shut-down commands, or raises control concerns) and "covered AI vulnerabilities" (weaknesses that could compromise safety or mission performance). To encourage transparency, the bill protects individuals who report these issues in good faith from professional retaliation or negative contract actions, and it safeguards sensitive or classified information. A designated official will categorize these reports to determine whether they require a local, program-level, or Department-wide response, with the most serious issues requiring a formal corrective action plan and verified fixes before the AI system can be used again. Finally, the Secretary of Defense must submit annual reports to Congress between 2027 and 2031 detailing the program's findings, systemic trends, and specific summaries of any incidents that resulted in bodily harm or death to members of the Armed Forces.