This Algorithm Can Tell When You’re Lying in an Email

Here’s an update that might make you happy or worried: a team of researchers from London has developed an algorithm that can detect lies in emails.

The team studied an archive of emails in order to understand the multifaceted aspects of digital language. This included the micro level (the ability of word use), macro level (message development) and meta level (intertextual exchange cues) of digital language. What they found is quite interesting but also pretty straightforward.

It appears that deceitful emailers over-structure their arguments, include quite a lot of flattery, and avoid the use of personal pronouns. They also “imitate” the linguistic style of the recipient because they want to appear more accommodating.

This approach comes from big data, as it combines statistics with natural language processing patterns that tell us when someone’s being deceitful.

The algorithm has been designed to protect organizations against fraud and financial loss and could help companies and organizations train their managers to identify linguistic cues to spot deceit and dishonesty. In other words, this lie-detection algorithm could save organizations a lot of money – it could also tell your boss whether you’re really sick or you’re just lying (in your email).

Advertisement