The future of lying: How technology helps filter the facts

We’re constantly coming up with new ways to lie. Fake reviews, “astroturfing” (concerted efforts at online propaganda), and even everyday lines like “On my way” or “Sorry, phone was dead” have all become routine forms of deception in today’s hyper-connected world.

Out of every form of online communication, email to people we know has been found to be the most honest medium. We’re somehow aware that it leaves a permanent record – that text-based lies are probably a bad idea because they can come back to bite us.

This is a relatively new phenomenon. The earliest evidence of writing is about 5,000 years old. In fact, most of our evolution as a species unfolded without any record of communication. You’ve probably written more in the last 24 hours than all of human prehistory.

These digital records change how we think about trust and deception. LinkedIn profiles, for example, have been found to be more accurate than “offline” résumés. Our ability to judge the personality of a good friend, meanwhile, is as reliable as judging a stranger’s Facebook page.

A “network effect” is what keeps deception in check. Even with online dating, men typically lie about their height by just nine-tenths of an inch. Why not more? Because they’re hoping to meet the person, which would expose an obvious lie and jeopardize any chance of connection.

In 50 years of research into our ability to detect deception, our most powerful tool for spotting lies is our innate bias towards truth. Giving someone the benefit of the doubt is our default setting.

There are no tell-tale signs of deception. It doesn’t matter whether we analyze body language, study facial expressions, or look someone in the eye. Studies have shown, time and time again, that there just isn’t one reliable “smoking gun” when it comes to lying.

Our average success rate of detecting deception is 54%. Even then, context plays a huge part. If smuggling something through customs, for example, our pupils might dilate or our voice may change pitch – but these things still cannot be accurately judged by humans.

Algorithms are better than us at spotting fake hotel reviews (80% success rate). The main giveaway is the use of prepositions relating to space, like bathroom size. You need to have been there to know that. (Liars tend to rely on narratives and use the first-person.)

In all other situations, algorithms aren’t so effective – yet. The closest thing we have to a built-in lie detector is the same approach made famous by TV detective Columbo: to question everything until the conclusion is clear. For now, technology is just a tool to help us do that.