The U.S. Department of Defense is requesting $30.3 million over five years for a program called «Polygraph+» (also referred to as Polygraph Next), which aims to modernize lie detection through artificial intelligence, machine learning scoring algorithms, and «standoff sensing»—a technique that captures physiological readings without physical contact with the subject.
According to budget documents first reported by [source], the initiative seeks to «modernize federal polygraph and credibility assessment technologies» to enhance their accuracy and reliability. The program would be administered by the Defense Counterintelligence and Security Agency (DCSA), which handles background checks for the federal government. If approved by Congress, the technology would be deployed for vetting job applicants and «insider threat detection.»
However, critics argue this represents yet another misguided attempt to solve an inherently unsolvable problem. «It’s a misguided effort to reduce the complex to something that is tangible,» says Kyri Kotsoglou, a legal scholar at Northumbria University in the UK who researches polygraph use in justice systems.
The proposal emerges amid heightened internal tensions at the Pentagon. Under Defense Secretary Pete Hegseth, the department has increasingly relied on polygraph examinations to identify sources of alleged press leaks. In September, the New York Times reported that staff members on the Joint Staff underwent polygraph tests following news coverage about depleted U.S. weapons stockpiles in the Iran conflict.
While specific technologies remain unclear—and the DCSA did not respond to requests for details—previous Pentagon efforts provide clues. In 2023, the Defense Innovation Unit (DIU) solicited companies with deception detection products. Two were selected: Presage Technologies, which claims it can measure heart rate and breathing through standard cameras, and Altec Research, a medical sensor company expanding into non-contact sensing. A DIU-released video of Altec’s prototype shows it tracking head movement, facial skin temperature, and pore activity. Neither company responded to comment requests, and the DIU declined to comment.
Polygraph technology has scarcely evolved since its invention in the 1920s. Examiners still rely on blood pressure, pulse, respiration, and sweat measurements, comparing physiological responses to baseline questions («Is the sky blue?») against target questions («Have you ever committed a crime?»). The federal government conducts tens of thousands of these tests annually for employee screening, yet the technology’s reliability has been repeatedly questioned, and results are rarely admissible in court.
In 1983, Congress’s Office of Technology Assessment found minimal evidence supporting polygraph use for employee screening. Two decades later, a 2003 National Research Council (NRC) report characterized the evidence for its effectiveness as «weak at best.» Research suggests humans detect lies with barely better than chance accuracy without technical assistance. While the American Polygraph Association claims 80% to 94% accuracy, the NRC report noted that even such accuracy rates in screening could produce substantial errors—applied across the Department of Defense’s 2.8 million employees, an imperfect system could falsely accuse tens of thousands.
Additional concerns include subjective interpretation, with examiners—and people from minority groups are more likely to be judged as deceptive. Furthermore, trained individuals can employ countermeasures to defeat the test, such as artificially elevating physiological responses to baseline questions by stepping on a hidden pin.
«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 adds.
Various alternative lie detection strategies have been attempted over the years, from thermal cameras to pupil trackers to brain scans. None have produced reliable results outside laboratory settings. The fundamental problem persists: there is no universal physiological indicator of deception. «There is still no Pinocchio’s nose,» says van der Zee.
AI could theoretically improve detection by identifying patterns examiners miss. Algorithms might also enable «multi-modal» deception detection, combining multiple measurements into an overall deception score that’s harder to manipulate. According to van der Zee, three underlying processes occur during deception: physiological stress, cognitive load, and conscious efforts to conceal lying. 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 new—in the 2000s, researchers at Manchester Metropolitan University in the UK developed a system called [name] that generated deception scores from video footage, later incorporated into iBorderCtrl, an EU-funded pilot. In the U.S., a project called AVATAR combined eye tracking, voice analysis, and body movement detection for border crossings. All these initiatives quietly disappeared.
Kotsoglou argues that merging AI with polygraph technology represents «the worst of both worlds,» adding uncertainty atop invalidity. Even if AI or machine learning identifies previously unseen patterns in physiological data, it cannot reliably link them to deception 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 justice systems. She fears new lie detection forms will, like the polygraph, serve more as psychological props than scientific tools.
«It seems very much a response to the concern of the current administration to leaks and perceived lack of loyalty,» says Oswald. «[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.