Science News

Researchers Develop Lie Detection Software Using Court Case Data

Image via Pixabay

University of Michigan researchers have developed new lie-detecting software that has a 75 percent accuracy rate of catching someone in a fib. The software works by analyzing the words and gestures made by the speaker, and unlike a polygraph test, it can be used without the speaker’s knowledge.

Over 120 court cases were analyzed by the researchers, who were then able to identify signs of a lie by comparing certain notable behaviors to the actual verdict. To identify a liar, the authors said certain behaviors such as looking directly at the questioner, excessive hand movements, and greater attempts to sound certain are just a few of many that are found more commonly in liars than in those who are honest.

Lying individuals moved their hands more. They tried to sound more certain. And, somewhat counterintuitively, they looked their questioners in the eye a bit more often than those presumed to be telling the truth, among other behaviours.

Rada Mihalcea, the leader of the project, explained that “There are clues that humans give naturally when they are being deceptive, but we’re not paying close enough attention to pick them up. We’re not counting how many times a person says ‘I’ or looks up.”

Mihalcea added that humans are too focused on interpreting the meaning of what a person is saying to catch the subtle indicators of a lie.

Right now, for the software to work, an observer must manually enter the behaviors exhibited by an individual being monitored for truthfulness. They are in the process of automating this functionality within the software itself.

The team plans on advancing the software’s detection rate further by adding more metrics. By using thermal imaging, they want to track a person’s heart rate, respiration rate and any fluctuations in their body’s temperature — all of which could be studied without physically touching the individual.

The researchers did acknowledge that because the verdict of the trials were used to indicate whether a person was actually lying or not, any erroneous verdicts could have directly impacted accuracy of their software with false positives or false negatives.

The software already is comparable in accuracy to that of a polygraph test, which holds an accuracy rate of anywhere between 70 and 90 percent. Polygraphs are much more intrusive, however, and require the participation of an individual in order to work. If the researchers continue to add new metrics and better functionality to the software, it may prove to be a much more viable tool for law enforcement, juries, and mental health professionals to determine honesty.

Click to comment
To Top

Hi - Get Important Content Like This Delivered Directly To You

Get important content and more delivered to you once or twice a week.

We don't want an impostor using your email address so please look for an email from us and click the link to confirm your email address.