--- title: "This Startup Says It Can Tell From Your Eyes If You’re Lying" description: "A startup called Converus created a lie detector that uses machine learning to analyze tiny changes in pupil dilation. There's just one issue." date: "2018-12-05" modified: "2018-12-05" authors: - name: "Jon Christian" job_title: "Executive Editor" link: "https://futurism.com/authors/jonc" url: "https://futurism.com/the-byte/lie-detector-eyedetect" categories: - "Artificial Intelligence" tags: - "converus" - "eyedetect" - "lie detector" - "the digest" --- # This Startup Says It Can Tell From Your Eyes If You’re Lying ![A startup called Converus uses machine learning to analyze tiny changes in your pupil dilation to detect when you're lying. There's just one issue.]() *\Image: Tag Hartman-Simkins\* ## Everybody Lies It's old news that polygraph tests, which are supposed to detect whether someone is lying based on their physiological responses during questioning, are [unreliable and prone to bias](). Now, a startup called Converus says it's invented a better alternative: EyeDetect. The system uses machine learning to analyze tiny changes in a subject's pupil dilation and reaction time, which lets it guess whether you're fibbing. It's cheaper than a polygraph, the company says in a [new *Wired* feature](), and more accurate. ## Lie to Me It's illegal for most private employers in the United States to use lie detectors on their workers, but EyeDetect has rapidly caught on with the U.S. government and in other countries. The Department of State contracted Converus to help it hire workers in Guatemala, and the U.S. Embassy in Paraguay has used EyeDetect, *Wired* found. The Defense Intelligence Agency and U.S. Customs and Border Protection are conducting trials with the technology, too. Converus also says that Best Western, FedEx, McDonald's, and other well-known companies are all clients, though none confirmed that claim with *Wired*. ## Liar Liar There's just one problem: experts aren't sure EyeDetect is any better than the notoriously unreliable polygraph test. The company says it's 86 percent accurate, *Wired* reported, but no researchers outside the company have conducted any peer-reviewed studies on it — and some of Converus's own studies have found accuracy rates as low as 50 percent. "It's astounding to me that there are paying customers deploying this technology and actually screening people with it," University of Minnesota professor of psychology and neuroscience William Iacono told *Wired*. If those concerns are true, the story is less about ingenious machine learning and more a cautionary tale about using artificial intelligence buzzwords to foist an unreliable test on companies — and innocent people — whose lives could be upended by its results. **READ MORE:** [An Eye-Scanning Lie Detector Is Forging a Dystopian Future]() \[*Wired*\] ***More on lie detectors:*** *Don’t Even TRY Lying to This Polygraph Test — Your Micro Emotions Will Give You Away* ## Author I'm responsible for editing, assigning, and scouting at Futurism, as well as occasionally writing for the site. That work involves keeping an eye on a wide range of narratives and issues, but over the past few years I've become increasingly interested in how AI is shaping the future of media, the web, and information itself. My day-to-day often involves collaborating on reporting and commentary projects bylined by my colleagues, but I try to find time to do my own reporting as well; stories I've broken for Futurism have been cited by publications including the New York Times, the Washington Post, the New Yorker, Wired, Ars Technica, New York Magazine, the Columbia Journalism Review, Bloomberg, and more. I grew up in Southern Vermont and attended Vanderbilt University. Prior to Futurism, I contributed to outlets including the Boston Globe, Vice, Wired, Slate, the Atlantic, the Outline, Ars Technica and more, and did stints in farming and food service. I currently live in Brooklyn, New York, and in my free time I enjoy pinball, word games, cycling, running, and music production. ### Author social links [Bluesky]()