The U.S. Department of Defense has requested $30.3 million over five years to develop an upgraded lie detection system, according to budget documents. The initiative, dubbed «Polygraph+» or «Polygraph Next,» aims to modernize federal polygraph and credibility assessment technologies through artificial intelligence and machine learning scoring algorithms, along with a technique called «standoff sensing»—the ability to capture physiological readings without physical contact with the subject.
The program would be managed by the Defense Counterintelligence and Security Agency (DCSA), which handles background investigations for the federal government. According to the budget proposal, which Congress has not yet approved, the technology would be deployed for vetting job candidates and «insider threat detection.» Specific technologies remain unspecified, and the DCSA did not respond to inquiries for additional details.
This effort arrives amid heightened internal tensions at the Pentagon. Under Defense Secretary Pete Hegseth, the department has increasingly employed polygraph examinations in attempts to identify sources of alleged media leaks. In September, the New York Times reported that personnel on the Joint Staff underwent polygraph testing following news coverage about depleted U.S. weapons stockpiles during the conflict with Iran.
Prior Pentagon initiatives may offer hints about the direction of Polygraph+. In 2023, the Defense Innovation Unit (DIU) launched an open call for companies developing deception detection products. Two firms were selected: Presage Technologies, which claims its standard cameras can measure heart rate and breathing rate, and Altec Research, a medical sensor company expanding into non-contact sensing. A video of Altec’s prototype released by the DIU demonstrates tracking of head movement, facial skin temperature, and pore activity. Neither company responded to requests for comment, and the DIU declined to comment.
Traditional lie detection technology has remained largely unchanged since the polygraph was invented. Examiners measure blood pressure, pulse, respiration, and sweat to assess truthfulness, comparing physiological responses to baseline questions like «Is the sky blue?» against target questions like «Have you ever committed a crime?» The federal government administers tens of thousands of these tests annually during employee screening, yet the technology’s reliability has faced repeated challenges, and results are rarely admissible in court.
In 1983, Congress’s Office of Technology Assessment concluded there was very limited evidence supporting polygraph use for employee screening. Two decades later, in 2003, the U.S. National Research Council (NRC) characterized the evidence for its effectiveness as «weak at best.» Research indicates humans can detect lies with slightly better than chance accuracy without technical assistance. While the American Polygraph Association claims 80% to 94% accuracy, the 2003 NRC report noted that even a screening test at this accuracy level would generate numerous errors. With the Department of Defense employing 2.8 million people, an imperfect system at that scale could falsely accuse tens of thousands.
Additional concerns exist. Polygraph interpretations are frequently subjective, and individuals from minority groups face higher likelihood of being judged deceptive. Furthermore, with training, interviewees can learn countermeasures to defeat the test—for instance, artificially elevating physiological responses to baseline questions by stepping on a pin concealed in their shoe.
«If you know how it works, you can beat it,» says Sophie van der Zee, an associate professor studying deception at Erasmus University in Rotterdam. She notes the machine’s primary effect is deterrence—subjects often confess before testing begins. «But that only works if people think a polygraph works,» she observes.
Various alternative lie detection strategies have been attempted over the years, employing technologies from thermal cameras to pupil trackers to brain scans. None have produced reliable results outside laboratory settings. The fundamental problem: no single telltale sign of lying applies universally to everyone all the time. «There is still no Pinocchio’s nose,» van der Zee says.
AI could theoretically improve detection if it identifies patterns in data that human examiners miss. AI algorithms are also better suited for «multi-modal» deception detection, which combines multiple measurements into an overall deception score that’s harder to manipulate. According to van der Zee, three processes occur beneath the surface that lie detection attempts to capture: physiological stress, cognitive load, and conscious efforts to conceal deception. Current polygraph technology addresses only one.
«The more you can have combined methods that approach it from these three different angles, the more successful you will be,» van der Zee says. This isn’t a new concept—in the 2000s, researchers at Manchester Metropolitan University in the UK developed a system called Silent Talker that generated deception scores from video footage. This was later incorporated into iBorderCtrl, an EU-funded pilot program. In the U.S., a project called AVATAR integrated eye tracking, voice analysis, and body movement detection into a border crossing tool. All these initiatives quietly disappeared.
Kotsoglou describes combining AI with polygraph technology as «the worst of both worlds» because it adds uncertainty to an already invalid method. Even if AI or machine learning detects previously unseen patterns in physiological data, it cannot reliably link them to lying because no real ground truth exists.
«Even if you have all the records in the world from polygraph tests, you don’t know whether those polygraph tests are right or not,» says Marion Oswald, a law professor who has co-authored work with Kotsoglou on polygraph use in the justice system. She worries that new forms of lie detection will, like the polygraph, serve more as psychological props than scientific instruments.
«It seems very much a response to the concern of the current administration to leaks and perceived lack of loyalty,» Oswald says. «[Lie detection is] being used as a threat, to intimidate and force people to confess to things, as opposed to anything that’s actually getting valid information.