Your Computer Knows You Better Than Your Mom

Why machines can predict your personality more accurately than your family or friends

Max Power/Corbis

Think your loved ones know you better than anyone else? An algorithm might have them beat—a new study shows that computers are better at predicting our personalities than our family or friends.

An international team of psychologists and computer scientists pitted computers against human personality prediction skills in the study. Co-authors Youyou Wu of the University of Cambridge and Michal Kosinski of Stanford University used a database called myPersonality, which has conducted hundreds of thousands of online personality tests to date.

The authors used the myPersonality results to tie personality traits to Facebook likes, then developed an algorithm that could predict personalities based on likes alone. Finally, they compared the results of their predictive algorithm to personality analyses performed by friends and family members. The result? The algorithm consistently outperformed family members, acquaintances and even close friends.

But don't think you can throw off the algorithm's analysis by posting different stuff on your feed. Instead, lay off the like button. Though polls confirm that people put their best (and most boastful) face forward on social media, what you like can predict how you purchase, listen to music or browse the Web—a much more powerful predictor than what you say about yourself online.

So what do your likes really say about you? Plenty. The model found that Douglas Adams readers are more spontaneous, Gucci fans are more outgoing and “Deal or No Deal” watchers are more conventional and conservative. And with our digital footprints growing every time we click “like,” it’s only a matter of time before our computers know us better than we do. (They can already beat us at Texas Hold 'Em.)

Get the latest stories in your inbox every weekday.