Pentagon Seeks AI to Monitor Inmate Calls at Fort Leavenworth

The Defense Department is preparing to deploy artificial intelligence systems to monitor phone calls and messages at its military prison facilities in Fort Leavenworth, Kansas, according to a new contracting opportunity.

DOD is soliciting proposals from companies that can provide cloud-based AI monitoring tools capable of flagging potentially suspicious language in real-time communications from more than 650 inmates housed at the Military Correctional Complex. Submissions are due by August 13.

Semantic and Agentic AI Requirements

The Pentagon's request specifies two distinct AI capabilities. Officials want semantic AI that can interpret the context and meaning of phrases, along with agentic AI that can plan and make decisions based on prompts. A key requirement is the system's ability to identify when inmates use new slang terms in place of standard language to discuss criminal activities or drug names.

The technology must integrate with the MCC's existing ViaPath Phone system, which already provides real-time security monitoring. The new AI layer would add transcription capabilities and a web-based dashboard for tracking non-privileged communications.

Operational Scope and Features

The Military Correctional Complex houses military prisoners convicted under the Uniform Code of Military Justice across facilities of varying security levels. This includes the U.S. Disciplinary Barracks, the only maximum-security facility within DOD.

Beyond language detection, the AI system must offer near-instant Spanish-English translation and multi-language options. Officials expect the technology to generate law enforcement reports with evidentiary value, provide targeted monitoring during emergencies, and collaborate with MCC investigators during standard work hours.

The system must be certified for use on Amazon Web Services' government cloud infrastructure, indicating the Pentagon's preference for established cloud platforms with security clearances.

Why it matters

This procurement reflects a broader trend of law enforcement and correctional facilities turning to AI for surveillance operations. While officials frame the technology as enhancing staff and inmate safety, the deployment of AI systems that can interpret context, detect evolving slang, and make autonomous decisions in a prison setting raises questions about oversight, accuracy, and the potential for false positives that could affect inmate privileges or legal proceedings. The requirement for agentic AI—systems that can plan and make decisions—represents a more advanced application than simple keyword flagging.

Defense Scoop first reported the contracting opportunity and details of the Pentagon's AI monitoring plans.