The Pentagon wants $30 million to build an AI-powered lie detector

The Pentagon is asking for $30.3 million to be spent over the next five years on an improved form of lie detector, according to a Department of Defense budget request reported by MIT Technology Review on September 25. The figure and its purpose come from the department's own budget documents.

The program appears in those documents under two names: "Polygraph+" and "Polygraph Next". Both describe the same effort to bring new computing methods into the way polygraph results are judged, rather than making incremental changes to the existing process.

According to the report, the work has two stated focus areas. The first is scoring algorithms that use artificial intelligence and machine learning; the second is a technique the request calls "standoff sensing".

Traditional polygraph examination relies on contact sensors that record physiological signals, with a human examiner interpreting the output. Moving the scoring step to machine learning suggests the system could search for patterns associated with deception across more data and more dimensions than an examiner would normally weigh.

So far, the public record stops at the budget request. The report does not name a contractor, a deployment timeline, a testing location, or any technical performance target such as an accuracy figure.

The budget structure is worth noting. Spread across five years, $30.3 million works out to roughly $6 million a year — a pace that reads more like sustained research funding than a one-off procurement.

Polygraph technology has long been contested in legal and security settings, where its scientific reliability is the central objection. A version built around algorithmic scoring is likely to reopen the same questions about accuracy, explainability, and the limits of acceptable use.

What to watch next: whether Congress preserves the requested amount during budget review, and when the Defense Department publishes a more detailed program description.