The Pentagon requests $30.3 million over five years to develop an AI lie detector, outlining goals, funding phases and oversight amid privacy concerns.

The Pentagon has requested $30.3 million to develop an artificial‑intelligence lie detector over the next five years.

The money would be allocated through the department’s research budget and would fund prototype design, algorithm training and field testing. Officials said the aim is to create a system that can evaluate verbal and non‑verbal cues for signs of deception.

Funding request details

The proposal outlines a phased approach. The first year would focus on data collection and model development. Subsequent years would expand testing to controlled environments and refine the technology for operational use.

Program objectives

Program leaders describe three core goals. First, the system must process audio and visual inputs in real time. Second, it should generate a confidence score that indicates the likelihood of false statements. Third, the technology must meet existing security and privacy standards set by the department.

Contractors with expertise in machine learning and signal processing are expected to compete for development contracts. The Pentagon has not disclosed which firms will receive awards.

Critics of similar projects have warned that AI‑based deception analysis can produce false positives and raise civil‑rights concerns. Defense officials responded that the program will include rigorous validation and oversight before any deployment.

The request will be reviewed by congressional committees that oversee defense spending. If approved, the first prototype could be demonstrated by the end of the program’s third year.

Stakeholders from intelligence agencies, law‑enforcement bodies and allied militaries have expressed interest in the technology’s potential applications, though no formal agreements have been announced.