Algorithm Detects Lies That People Tell Online

Truth and LiesA new computer algorithm can find and detect the lies that people tell in their emails, social media or online dating profiles.

A group of four researchers at Cass Business School in the United Kingdom claim to have developed a lie-detecting algorithm that can find a person’s lies simply by analyzing their choice of words, sentence structure and context in electronic communications.

The algorithm was developed by anaylizing tens of thousands of archived business emails that compared truthful email content with email content that contained lies and falsehoods.

Computer analysis of these emails showed that people who lie tend to use more adjectives and descriptions in their communications than truthful people. Whereas truthful people tend to use more personal pronouns like “me”, “I” and “mine”.

The researchers believe that the reason for this is that people who lie may be trying to detach themselves personally from the false content in their communications, while also trying to obscure the meaning of what they are saying.

In addition, the research also found that people who lie tend to connect unrelated thoughts and sentences together, and mimic the sentences of the other person that they are having communications with.

The lie-detecting algorithm outperforms the average person at identifying lies. On average, people are able to identify a lie 54 percent of the time. However, the algorithm can detect a lie around 70 percent of the time. Studies show that people are really bad at recognizing the lies and falsehoods that other people tell.

The researchers developed the algorithm to help people and businesses protect themselves against email phishing, online dating scams and any other fraud that uses computer-based communications. They hope that someday their system will be completely automated to detect lies in every form of digital communications. But before they offer this digital lie detector commercially, they want to increase the system’s accuracy.

With improved lie-detection accuracy, this program could really be of use in helping people avoid email phishing and online dating scams. The personal and financial savings from an algorithm like this could be priceless for both people and businesses.