# Pentagon Seeks $30 Million for AI-Powered Lie Detection System
The U.S. Department of Defense requested $30.3 million over five years to develop an advanced lie detection system that combines artificial intelligence with novel sensing techniques. The program, dubbed Polygraph+ or Polygraph Next, aims to replace aging polygraph technology with machine learning algorithms and a method called "standoff sensing" that operates without physical contact.
The funding request reveals the Pentagon's push to modernize interrogation and credibility assessment tools used across military and intelligence agencies. Traditional polygraphs measure physiological responses like heart rate, blood pressure, and skin conductivity, but they generate high false-positive rates and face scientific scrutiny. The new system targets these weaknesses by leveraging AI to interpret biological signals more accurately.
Standoff sensing represents the technical shift at the program's core. Instead of attaching sensors to a subject's body, standoff techniques detect physiological indicators from a distance using optical, thermal, or other non-contact methods. This approach could reduce subject discomfort and potentially bypass countermeasures that defeat traditional polygraphs. The exact sensing modalities remain unspecified in the budget documents, though infrared and video-based biometric analysis are likely candidates.
AI and machine learning algorithms will score results from these sensors, identifying patterns that correlate with deception. The Pentagon likely seeks to improve on traditional threshold-based scoring systems by training models on large datasets of interrogation sessions, flagging behavioral and physiological markers humans might miss. This represents a substantial leap from the binary interpretations of conventional polygraphs.
The timing reflects broader Pentagon interest in AI-driven security tools. Defense agencies already deploy machine learning for threat detection, identity verification, and intelligence analysis. Polygraph+ fits this modernization agenda, particularly as the military manages classified information access and investigates counterintelligence threats.
However, the program faces genuine scientific obstacles. Polygraphy remains controversial in academia and courts. The National Academies of Sciences found in 2003 that polygraphs perform significantly better than chance but produce substantial error rates. No peer-reviewed evidence demonstrates that AI-enhanced versions overcome these limitations. Deception involves complex neurological responses without clear, universal physiological signatures. Machine learning models can amplify bias if trained datasets contain skewed interrogation outcomes.
Privacy and ethical concerns accompany the Pentagon's plan. Standoff sensing that identifies lies without consent raises legal questions around surveillance and due process. Military and intelligence use contexts differ from civilian law enforcement, where polygraphs already face admissibility challenges in court.
The five-year timeline suggests a phased approach. Early funding likely supports algorithm development and sensor integration. Later phases probably involve field testing across military interrogation facilities and intelligence community offices. Success depends on demonstrating false-positive and false-negative rates substantially lower than current polygraphs and gathering sufficient validated training data.
Intelligence analysts and military personnel conducting interrogations will ultimately decide whether Polygraph+ offers genuine improvement or simply repackages old problems in new technical language. The $30.3 million investment signals Pentagon confidence in the concept, but delivering reliable, unbiased deception detection remains an unsolved technical problem regardless of the funding level.