Updated
Updated · MIT Technology Review · Sep 25
Pentagon Seeks $30.3 Million for AI Lie Detector as Leak Hunts Expand
Updated
Updated · MIT Technology Review · Sep 25

Pentagon Seeks $30.3 Million for AI Lie Detector as Leak Hunts Expand

1 articles · Updated · MIT Technology Review · Sep 25

Summary

  • $30.3 million over five years is in the Pentagon budget request for “Polygraph+,” an AI- and machine-learning upgrade to federal lie detection that still awaits congressional approval.
  • The program would add algorithmic scoring and “standoff sensing” that captures physiological signals without attached devices, for employee vetting and insider-threat detection under the Defense Counterintelligence and Security Agency.
  • The push comes as the Defense Department intensifies leak investigations; the New York Times reported about 50 Joint Staff officers were polygraphed this month after stories on depleted US weapons stockpiles in the Iran war.
  • Past Pentagon work points to camera-based and non-contact sensing prototypes, but experts say lie detection lacks a reliable biological marker, making AI-enhanced systems vulnerable to the same false positives, bias and countermeasures as traditional polygraphs.
  • That skepticism is longstanding: federal polygraphs have been criticized for decades, and a flawed system applied across the DoD’s 2.8 million staff could wrongly implicate large numbers of people.

Insights

Can AI really spot a liar, or is the Pentagon just building a thirty million dollar high-tech intimidation tool?
Will invisible AI sensors scanning your face and heart rate become the new mandatory hurdle for federal employment?