Your Computer Understands Your Personality Better Than Your Friends Do

Whenever you use a computer, you leave behind a rich seam of personal information—and computers can use that to understand your personality far better than your family or friends.

Researchers from Stanford University in California and the University of Cambridge have performed a study which compared the personality traits of 86,000 participants acquired via a questionnaire with an analysis of their Facebook activity. Where the team had gathered data from over 100 Facebook likes for a person, their algorithms could determine a person's personality more accurately than their friends and family—and almost as well as their partner. The research is published in PNAS.

Interestingly, the techniques aren't limited to Facebook, so any analysis of other evidence left behind on a computer—the websites our visit, the purchases you make on Amazon, whatever—could all be used to contribute to working out what kind of person you are. It's not just idle research either: the team suggest that their insights could easily be used to predict your suitability for a job. Instead of applying for new employment, algorithms would spot from scouring internet data that you might be suitable—then headhunt you.

Scared? It's perhaps still a little early to panic. After all, this analysis is based on just five major personality traits, while many psychological analyses use hundreds of dimensions to build an accurate picture of you and your behaviour. Until computers can match that, no algorithm will ever know you as well as, say, your spouse. But regardless, the power of the algorithm is clear; it might be time to pay a little more care to what you like on Facebook.

[PNAS via New Scientist]