Pentagon requests $30 million for AI-powered lie detector program

By Splitfeed EditorialLast updated Sep 26, 2026, 19:17 UTC

On September 25, 2026, MIT Technology Review reported that the US Department of Defense had submitted a budget request for $30.3 million to develop an upgraded lie detection program over a five-year period. The program, referred to in the request as either Polygraph+ or Polygraph Next, would apply artificial intelligence and machine learning to the scoring algorithms that interpret physiological data. It would also invest in standoff sensing, a technique that reads physiological signals from a subject at a distance rather than through the wired contact methods traditional polygraphs require.

The significance of this request goes well beyond a routine procurement upgrade. Conventional polygraph technology has long been contested on scientific grounds, with researchers and courts repeatedly questioning whether physiological stress signals reliably indicate deception. Adding AI scoring layers to that foundation raises compounding concerns: if the underlying data correlation is disputed, then a machine learning model trained on that data may entrench rather than correct the problem. Standoff sensing adds a further dimension, because detecting signals without a subject's knowledge or explicit consent moves the technology closer to covert surveillance. The program sits at a crossroads between national security vetting, where the DoD has used polygraphs for decades, and a broader debate about whether AI can or should be used to make inferences about human mental states.

No reactions from the Pro-AI, Anti-AI, or Middle Ground camps have been published at this stage. Pro-AI voices would typically argue that machine learning algorithms, trained on larger and more diverse datasets than older scoring rubrics, could reduce human bias and improve consistency in an already-established vetting process. Anti-AI voices would be expected to raise alarms about the scientific validity problem being laundered through algorithmic authority, and about standoff sensing enabling monitoring without consent. A middle ground position would likely call for rigorous independent validation of any AI scoring system before deployment, and for clear legal boundaries around when standoff sensing could be used.

The detail most worth tracking is whether Congress approves the budget line as submitted or attaches conditions requiring scientific review or oversight mechanisms. Any testimony from independent researchers or civil liberties groups during appropriations hearings would also clarify how much institutional resistance the program faces before funding is confirmed.