Here’s how an AI Lie Detector can tell when you are fibbing


Nowadays it is difficult to figure out when a person is telling truth or not. It is a belief people who vow under oath, he will definitely tell the truth but it is not correct. He may also sometimes lies. As we all know AI ( Artificial Intelligence) is everywhere. If you are in a depression state, artificial intelligence can figure out with your voice easily. Have you wondered why we are talking about it?

This whole idea comes from an AI fib sniffer that is in news because of a security project that held in Europe called iBorderCtrl. It includes a technology known as “ deception detection”. Basically, it consists of a two-step process and a lie detection part.

The main thing about “deception detection” is that it analyses the micro – expressions of travelers to figure out if they are lying or not includes ‘lips protruded’, ‘eyebrows frown’, ‘ frequency of vocal patterns’ etc. It involves webcam to answer questions from a computer animated background.

It also shows the polygraph tests of troubling historic mind. I know it sounds like friction. But the main question comes that such an AI system is possible or not? There are various professors who worked on deception detection. One of them is “Rada Mihalcea”, a professor of computer science and engineering. It took almost a decade to know about how to construct one AI deception detector and also about its working.

Data plays a vital role while researches are done on the working of artificial intelligence and machine learning. They began with videos from actual cases to find out whether a defendant speaking in a trial in which they found guilty or not with the help of deception detector. The main aim of these videos does not just focus on courtroom videos but also predict deception.

According to her research, 121 video clips are used and had between a 60 to 75 percent accuracy rate. Some of the factors which she noticed that people who are not lying using the words such as b ‘I’, ‘We’ or ‘themselves’ whereas people who are lying using the words such as ‘you’, ‘yours’ or ‘they’. Not only this, stronger words that reflect certainty are used by the people who are telling a lie.

We can trust those people who use such words like ‘maybe’ or ‘probably’. People who are not lying would try to make up for the lie they are putting forward and try to seem more certain of themselves. They use both hands instead of just one hand as they are trying to convince others. They look directly into the eyes of the person questioning him. Not only this, other researchers whose name is Chatilla says that it will three to four years by AI that detects deception flawlessly by reading the emotions behind human expressions.

Comment here