Liar, liar! ‘Reading’ faces, Israeli tech spots fibbers with 73% accuracy

Israeli scientists say they have found a way to “read” minuscule movements in the face in order to spot fibbers, and have done so with 73 percent accuracy.

With highly sensitive electrodes placed to detect the smallest of movements by facial muscles, the researchers got their subjects to either speak truthfully or lie.

They fed details on the patterns of those facial movements into an artificial intelligence tool, and taught it to determine whether other people are lying or telling the truth.

Now, they are aiming to teach the AI tool to analyze face movements without electrodes. Instead, they want to develop the tech to follow faces in order to determine truthfulness via cameras — which could enable them to spot a liar from dozens of meters away.

“The accuracy of our test will rise further as we develop it more, and our hope is that eventually, after development and thorough testing, this could provide a serious alternative to polygraph tests,” Prof. Dino Levy of Tel Aviv University, part of the team that is developing the technology, told The Times of Israel.

“In the bank, in police interrogations, at the airport, or in online job interviews, high-resolution cameras trained to identify movements of facial muscles will be able to tell truthful statements from lies. Right now, our team’s task is to complete the experimental stage, train our algorithms and do away with the electrodes. Once the technology has been perfected, we expect it to have numerous, highly diverse applications.”

The research has been peer-reviewed and published in the journal Brain and Behavior. Levy said he believes that the new approach is less circumventable than polygraphs, in which some people can hide lying by regulating heartrate, blood pressure and respiration.

Illustrative image: scientists hope their lie-detecting system will eventually work via cameras (iStock vie Getty Images)

“Its far better than the regular polygraph line detector tests, which which people can fool,” said Levy. “Our test is a lot harder to track, as it’s based on changes in muscles that we’re not even aware of.”

Levy and his collaborators, including Prof. Yael Hanein and Liz Izakon, stress that the connection between facial muscle movements and truth was previously known. Their advance is detecting them with a high level of sensitivity, thanks to special electrodes invented by Levy, and curating an advanced AI tool that allows them to use the data they gathered to test others’ veracity.

Levy said: “We knew before now that facial expressions that are manifested by contractions in face muscles represent various emotions. This has been known for many years.

“But up until now when people tried to identify these small movements in face muscles, we can’t do — our brains and our perception aren’t fast or sophisticated enough to pick up these tiny movements in the face.

“Many studies have shown that it’s almost impossible for us to tell when someone is lying to us. Even experts, such as police interrogators, do only a little better than the rest of us. Our approach can effect a big change, and as accuracy levels rise further, become important technology.”