The Pentagon is seeking $30.3 million over the next five years to develop an upgraded lie detection system built around artificial intelligence, machine learning and a technique known as standoff sensing, according to a Department of Defense budget request. The initiative, called Polygraph+ or Polygraph Next, reflects a broader military and intelligence interest in automating judgment-heavy screening tools that have long been criticized for inconsistency and limited scientific reliability.
AI Screening Push
The proposal places new technology at the center of a process that has traditionally relied on human examiners interpreting physiological responses during polygraph tests. By introducing scoring algorithms trained with AI and machine learning, the Pentagon appears to be aiming for a system that can analyze patterns more quickly, more consistently and, in theory, with less examiner bias. The budget request suggests the department wants to move beyond a conventional polygraph model and toward a more data-driven screening framework.
That ambition comes with significant technical and policy questions. Polygraph testing has never been universally accepted as a definitive measure of truthfulness, and its use has often been confined to security screening rather than courtroom proof. An AI-enhanced version would not eliminate those concerns; instead, it would likely intensify scrutiny over how the algorithms are trained, what data they rely on and whether they can be independently validated. In national security settings, where false positives can affect careers and false negatives can create security risks, the margin for error is especially consequential.
The Pentagon's interest also fits a wider pattern in which defense agencies are exploring machine learning for surveillance, screening and decision support. Yet lie detection is one of the most sensitive applications imaginable. Unlike battlefield targeting or logistics optimization, it touches directly on trust, privacy and due process. Any system that claims to infer deception from behavior, physiology or remote signals will face intense debate from scientists, civil liberties groups and lawmakers.
Standoff Sensing Ambitions
A notable element of the program is its focus on standoff sensing, a term that generally refers to detecting or measuring signals from a distance rather than through direct contact. In the context of lie detection, that could mean using sensors to capture subtle physiological or behavioral indicators without the need for traditional wired attachments or close physical interaction. If successful, such a capability could make screening faster and less intrusive, while also broadening the settings in which it can be used.
But standoff sensing is also where the science becomes most contested. Remote detection of deception has long been a goal of researchers, yet the evidence base for reliable, real-world performance remains thin. Human stress, fear, cultural differences and medical conditions can all distort the signals that lie detection systems attempt to interpret. Adding AI may improve pattern recognition, but it does not automatically solve the underlying problem of distinguishing deception from other forms of arousal or anxiety.
The budget request indicates the Pentagon wants to invest in scoring algorithms as well as sensing methods, suggesting the program is intended to be both a hardware and software effort. That combination could make Polygraph+ more adaptable than older systems, but it also raises the stakes for oversight. If the technology is deployed in security vetting, personnel screening or intelligence contexts, the consequences of an opaque algorithmic score could be substantial.
Reliability Under Scrutiny
The broader significance of the request lies in what it says about the government's appetite for AI in high-stakes judgment systems. Defense officials are under pressure to improve screening and counterintelligence tools, especially as digital systems generate more data and adversaries become more sophisticated. Yet the history of lie detection is a cautionary tale: tools marketed as objective often prove far messier in practice.
A five-year, $30.3 million request is modest by Pentagon standards, but the program could still shape future procurement and research priorities. If Polygraph+ advances, it may become a test case for whether AI can meaningfully improve a field that has resisted definitive scientific validation for decades. If it fails, it may reinforce the argument that deception is too complex to be reduced to a score.
For now, the request shows that the Pentagon is willing to keep investing in the search for a better lie detector, even as the underlying science remains unsettled. The question is not only whether AI can make polygraphing more accurate, but whether a more advanced system can overcome the fundamental limits that have shadowed lie detection from the start.