Most people aren’t very good at knowing when they are being lied to. Study after study has shown that most people—including police officers, CIA and FBI agents, therapists, and even lawyers—might just as well flip a coin as trust their instincts about whether someone is being honest. In their 1991 paper “Who Can Catch a Liar,” two leaders in the study of lie detection, Paul Ekman and Maureen O’Sullivan, summarized the academic literature: “Average accuracy in detecting deceit has rarely been above 60%.”
Until O’Sullivan’s death in 2010, Ekman and O’Sullivan continued to study lie detection, and they discovered that a tiny percentage of people are “Truth Wizards,” who can identify liars with more than 80% accuracy. After testing approximately 20,000 people, Ekman and O’Sullivan identified fifty such “Wizards,” about one quarter of one percent of the tested population.
The Truth Wizard test was simple: People were shown videos of people and had to determine whether they were lying. According to O’Sullivan, there are two categories of clues to detect lies: thinking clues and emotional clues.
Thinking clues—such as stuttering, speaking slowly, and weaving stories with logical inconsistencies—are behaviors that arise because it is more mentally taxing to lie than to tell the truth. Emotional clues reveal themselves when people exhibit “micro-expressions,” facial tics so fleeting (1/15th to 1/25th of a second) that they are almost imperceptible. Recognizing a micro-expression of anger, happiness, fear, or disgust does not, by itself, tell you whether a person is lying. The trick to Truth Wizardry is to notice not only the micro-expression, but also the disconnect between what a person is saying and what flashes across his or her face.
For example, at a filmed press conference, Kim Philby, a British intelligence officer who was on the verge of being exposed as a double agent, was asked whether he was satisfied that the Prime Minister did not think he was the “third man” who assisted two other known Soviet spies. Philby confidently answered that he was. To a regular observer, Philby’s expression was perfectly consistent with his answer. But when a person is attuned to micro-expressions, it’s possible to see two smiles flash across Philby’s face. In an interview with Malcolm Gladwell, Ekman explained that “Twice, after being asked serious questions about whether he’s committed treason, he’s going to smirk. He looks like the cat who ate the canary.” Ekman recognized this micro-expression as “duping delight—the thrill you get from fooling other people.” Later in the interview, Philby was asked, point blank, whether he was the “third man.” Though he confidently denied it, his eyebrows shot up in a micro-expression of fear. See the video for yourself here.
Despite the obvious benefits of truth detection in police work and the judicial system, the only branch of law enforcement whose members are better than the general population at detecting liars is the Secret Service. Nonetheless, Ekman believes that—even if they can’t become Truth Wizards—with training all people can improve their ability to detect lies.
Can computers do any better? One relatively new technology is the Functional Magnetic Resonance Imaging test, or fMRI, which scans the flow of blood in the brain to produce a map showing which regions of the brain become active in response to a given stimulus. Some research has shown that certain regions of the brain are more active when we lie than when we tell the truth, and now companies like Cephos in Massachusetts and No Lie MRI in California market themselves essentially as offering the ultimate in lie detection: a full-color photo of a lie moving through the brain.
But last year, Justice Robert Miller in Brooklyn became the first judge in the nation to publish an opinion considering whether or not an fMRI test could be admitted as evidence to bolster the credibility of a witness. Justice Miller decided that the scientific community does not accept fMRI lie-detector tests as reliable, and, likening the fMRI test to an old-fashioned polygraph test, he excluded it from evidence.
A foreign court, considering another new technology—the Brain Electrical Oscillations Signature test, or BEOS—has come out differently. The premise of BEOS is that an electro-encephalograph can detect the difference in brain activity between when a human remembers an event, and when that human hears of an event for the first time. So in theory, when prosecutors read details of a crime to a suspect, BEOS should be able to determine whether the suspect has any memories of having committed it. In 2008, a court in India made news by sentencing a 24-year-old student to life imprisonment for murdering a former lover, based in part on a BEOS test showing the student had memories of lacing the dead man’s dessert with arsenic.
Currently, neither Truth Wizards nor truth machines are used in American courts to determine the veracity of a witness. But as long as psychology and computer science continue to produce new techniques, there will be pressure to try them out in front of juries. We at MB will keep you posted.