The Pentagon requested 30 million dollars to develop AI lie detectors

The US Department of Defense is planning to invest 30.3 million dollars over five years into a new project called Polygraph+ or Polygraph Next. This effort intends to update how the federal government verifies the truthfulness of employees and assesses internal threats.

The current polygraph system relies on measuring basic physical responses like pulse and breathing. These methods have remained largely unchanged since the 1920s and face long-standing questions regarding their scientific reliability.

Proposed technology would likely use sensors that do not require physical contact, such as cameras that track skin temperature or eye movement. These tools would then feed data into AI algorithms designed to calculate an overall deception score.

The project is managed by the Defense Counterintelligence and Security Agency. While the DOD has not confirmed the exact technical specifications, it has previously explored prototypes from companies like Presage Technologies and Altec Research.

Legal scholars and experts argue that adding AI to polygraphs could increase uncertainty rather than accuracy. Because there is no objective way to confirm if a test result is correct, an algorithm might simply learn to mirror flawed human judgments.

If you could not trust the machine in front of you, would you feel more or less likely to tell the truth?

This project remains in the budget proposal stage and has not yet been approved by Congress. Current research does not provide a proven way to detect lies through physical sensors alone, as there is no consistent physiological sign of deception that applies to everyone.