The Pentagon Wants $30 Million to Build an AI-Powered Lie Detector
Polygraph+ will use new sensor technologies and AI to vet staff and hunt leakers. Experts say AI-enabled lie detection is 'the worst of both worlds.'
By Amit Katwala | September 25, 2026
The Pentagon is seeking $30 million to develop an AI-powered lie detection system known as Polygraph+, according to a report from MIT Technology Review. The system would combine advanced sensor technologies with artificial intelligence to screen personnel and identify individuals leaking classified information.
The initiative reflects a growing push within the U.S. Department of Defense to modernize its vetting and counterintelligence capabilities. Traditional polygraph tests, long used in security clearance processes, have faced decades of criticism for their questionable reliability. Polygraph+ aims to address those limitations by integrating AI-driven analysis—but experts warn the approach may compound problems rather than solve them.
'The Worst of Both Worlds'
Critics argue that AI-enabled lie detection inherits the flaws of traditional polygraph testing while introducing new risks associated with algorithmic decision-making. Polygraphs measure physiological responses such as heart rate, blood pressure, and perspiration, but these signals are not reliable indicators of deception. Adding AI to the equation, experts say, does not fix the underlying science—it simply automates and scales a flawed method.
One expert described the combination as "the worst of both worlds," noting that AI systems can amplify bias and create a false sense of objectivity. Unlike traditional polygraph exams, which involve a trained human examiner, AI-driven systems may offer less transparency and fewer opportunities for appeal.
New Sensors, Old Problems
The Pentagon's Polygraph+ program promises to leverage cutting-edge sensor technologies that go beyond conventional polygraph inputs. While the specific sensors have not been detailed, potential options could include eye tracking, voice stress analysis, thermal imaging, and brain-computer interfaces. Each of these technologies has been studied for deception detection, but none has demonstrated consistently reliable results in peer-reviewed research.
AI could theoretically improve pattern recognition across multiple sensor streams, but without a validated ground truth for deception, the system risks producing confident but inaccurate assessments. In security contexts, false positives could ruin careers and erode trust within the workforce, while false negatives could allow genuine leaks to go undetected.
Implications for National Security and Civil Liberties
The Pentagon's interest in AI-powered lie detection raises broader questions about the role of automated surveillance in government and military settings. As AI tools become more prevalent in 2026, concerns about privacy, due process, and the ethical use of emerging technologies continue to grow.
Security experts and civil liberties advocates are likely to scrutinize the Polygraph+ initiative closely. If deployed, the system could set a precedent for AI-based credibility assessment across federal agencies and even the private sector. Whether the technology can meet its promises—or whether it becomes another cautionary tale in the rush to apply AI to complex human problems—remains an open question.
This article was originally published by MIT Technology Review. Image: Stephanie Arnett/MIT Technology Review | Adobe Stock, Rawpixel