Since the beginning of time, it has been difficult to discern when someone is lying. Now, artificial intelligence may help scientists get closer to this goal.
A team of researchers from Tel Aviv University installed sensors on the faces of volunteers to detect subtle movements as they tell lies and truths.
This system could detect lies with 73% accuracy. It is slightly lower than polygraph tests, which are accurate 80 percent of the time.
But, they claim that the current stage of development is only the beginning and that things will get better in the future.
The future will see AI-equipped cameras being used in interviews, online jobs and in police suspect interviews to determine if someone is lying.

It has been decades-long goal to be able to discern when someone lies. A team of Israeli scientists believes they are getting closer to that goal, using artificial intelligence.
Research has shown that people can distinguish lies from truth approximately 55% of the times. A polygraph machine, however, is accurate to up to 82%.
The results of the polygraph test are not sufficient to prove a person’s identity. Therefore, researchers around the globe have been working hard on innovative solutions.
Tel Aviv’s project used machine learning and artificial Intelligence to analyse muscle movement changes during lies.
This included small movements of the cheek muscles and of the eyebrows.
The stickers were printed on smooth surfaces and contained electrodes that can measure nerves and muscles.
Professor Levy stated that it was almost impossible to discern when someone lies to you. Experts, like police interrogators do not perform as well as the rest.
‘Existing lie detectors are so unreliable that their results are not admissible as evidence in courts of law – because just about anyone can learn how to control their pulse and deceive the machine.
‘Consequently, there is a great need for a more accurate deception-identifying technology.
“Our study assumes that our facial muscles contract when we lie down, but that no electrodes are sensitive enough to detect these contortions.”
Researchers attached stickers to cheek muscles just below the lip and above the eyebrows.
The volunteers were then asked to sit together in pairs, facing one another. After that, the words “line” or “tree” were said into their ears via headphones.
Participants were told to lie to or reveal the truth about the word that they’d heard to their partner.

The stickers were attached to the facial muscles by attaching special electrodes. These electrodes are above the eyebrow and the lip. The volunteers then said whether the truth was true or false.
The wearer was lying when he said ‘tree’ and heard ‘line’.
The results showed that participants could not detect the lies of their partners with statistical significance.
However, 73% of all lies can be detected by electrical signals from the stickers.
“Since it was an original study, the lie itself wasn’t complicated.” Professor Levy said that typically, when we lie, in real life we tell a longer narrative which contains both deceptive as truthful elements.
“In our research, we could hear the voices of the participants through their headsets. We also knew when they were lying.
“Hence, we used advanced machine learning techniques to train our program so that it could identify lies using EMG (electromyography signals) signals emanating from the electrodes.
“Using this method we were able to achieve 73% accuracy – while not perfect but better than other technologies.
“Another fascinating discovery is that different people lie using different facial muscles. Some lie with their cheek muscles, while others lie with their eyebrows.
According to the researchers, AI has ‘dramatic consequences in many spheres our lives’.

The future AI-equipped cameras will be used in online interviews, at airports and in police suspect interviews to check for fraud.
Professor Levy believes that electrodes could be replaced by video software, which can detect lies through facial muscles movements.
He stated that high-resolution cameras, which are trained to detect facial movements and distinguish truth from lies, can be used in the workplace, the police station, the airport or online for job interviews.
“Right now our task is to finish the experimental stage and train our algorithms before we get rid of the electrodes. We expect the technology to be used in many other applications once it is perfected.
The journal Brain and Behaviour published the paper.
In the past, critics have criticized artificial intelligence as a way to identify liars.
Manchester Metropolitan University, for example, suggested that AI could detect whether people lie at the border by analysing microgestures.
Ray Bull from the University of Derby’s criminal investigation department said, however, that the project is not credible because it does not show how microgeography can be used to detect lying.
He said, “They deceive themselves into believing it will ever prove to be substantial effective” and that they were wasting lots of money.
“The technology was based upon a fundamental misunderstanding of how humans behave when truthful and deceptive.”