The Department of Defense (DoD) has proposed a $30.3 million, five-year initiative to develop an AI-assisted lie detection system, referred to as Polygraph+ or Polygraph Next. The funding request appears in the Defense Counterintelligence and Security Agency's (DCSA) 2027 budget submission. If approved by Congress, the program aims to modernize federal polygraph and credibility assessment technologies, primarily for vetting prospective employees and detecting insider threats within the federal government.
The proposed system would incorporate artificial intelligence and machine learning algorithms to generate credibility scores. A key component of Polygraph+ is "standoff sensing," a technique that would allow physiological readings to be taken from individuals without requiring physical contact or attached devices. While the budget request does not specify the exact sensors to be used, earlier prototypes from the Defense Innovation Unit (DIU) in 2023 explored camera-based systems. These prototypes, developed by companies like Presage Technologies and Altec Research, measured metrics such as heart rate, breathing, head movement, facial skin temperature, and pore activity using standard cameras.
The DCSA, responsible for federal background checks, would operate the Polygraph+ system. The budget document outlines a need for automated scoring, decision aids, and cloud-based storage and analytics tools to refine the software and identify new diagnostic features. The initiative comes as the Pentagon has increased its use of traditional polygraph tests, particularly in response to investigations into alleged leaks of classified information. For instance, in September, approximately 50 Joint Staff officers reportedly underwent polygraph examinations following news coverage of depleted US weapons stockpiles.
However, the scientific validity of polygraph tests has faced scrutiny for decades. A 2003 review by the US National Research Council described the evidence for polygraph accuracy as "weak at best." Critics argue that traditional polygraphs, which measure physiological responses like breathing, pulse, blood pressure, and sweat, do not directly detect lies but rather indicators of stress or fear. Legal scholars and deception researchers have expressed concerns that integrating AI into such a system could introduce additional uncertainties, potentially exacerbating existing flaws rather than resolving them. Experts note that no reliable physiological indicator of deception has been consistently identified despite extensive research efforts.
The Defense Innovation Unit had previously issued a solicitation for its "Polygraph+" program in May 2023, seeking commercial solutions to automate and improve credibility assessments. The DIU's objective was to find technologies that could measure physiological and behavioral signals with non-contact sensors, aiming for less intrusive and potentially more reliable methods than existing polygraph techniques. The current budget request for Polygraph+ is pending congressional approval.