How Facebook Knows You Better Than Anybody Else

Who knows you in-and-out? Your family? Friends? Your spouse, maybe?

Who knows you in-and-out? Your family? Friends? Your spouse, maybe?

Some of you might believe that it’s Google (considering how much we rely on it for searches, navigation, via our Gmail etc). Well, It might actually be Facebook!

It seems that Facebook can predict how you’ll react to good or bad news, just as accurately as any of the above mentioned people. It may even know what your favourite cuisine is!

Researchers at the University of Cambridge and Stanford University studied how Likes on Facebook matched up with peoples’ (let’s call them Subjects) own answers on personality tests, as well as of those of their family or close friends’.

Inspired by the movie ‘Her’, researchers at these universities were curious to know how computers would do in evaluating personal traits. About 86,000 people were asked to complete a 100-question personality survey that helped the researchers stack personality traits.

Once the program they made had a database comprising of a sufficient amount of “likes” of objects, brands, people, music or books, the program was better at predicting a Subject’s personality than most people close to him or her.

The researchers then analyzed the respondents’ Facebook Likes to generate a model in which Likes were linked to these traits. Upon completion of such repeated data, these researchers deduced that the more Likes a person had on Facebook, the more accurate the computer’s results were.

Would you believe, it took only 10 Likes for the computer to outperform a work colleague, and 70 to do better than a close friend and 150 to outscore a family member, in their knowledge and prediction of the Subject’s reaction, preferences and many personal traits.

We all know that people are pretty good at predicting each others’ personalities as that is one of the most pivotal part of our interactions, but what’s amazing is the fact that computers can determine a person’s personality quicker and more accurately than his friends or family, by using just a single kind of digital data (Facebook Likes). Truly eye-popping!

There was a good explanation provided for this phenomenon: computers can do what they do and the way they do it because, they take all data at face value and treat the data equally, without any other criterion in their ‘mind’.
 Whereas people like you and me tend to forget, overlook or play down certain important pieces of information when evaluating someone’s personality traits. We also tend to give more weight to events that might have been fairly recent or more memorable, which leads to a biased result.

Computers on the other hand have the inbuilt ability to treat each piece of information objectively.

So are we at the Minority Report stage yet?

No. Despite all this research and advancement, this digital approach isn’t always entirely accurate. It can’t account for changes in people’s moods and behaviours. And given the fact that we humans are notoriously dynamic, this could be a problem.

Michal Konsinski, from Stanford’s Department of Computer Science thinks that this kind of computer modeling could help processes like career planning and job recruitment.

Think about it — People just entering the job market could benefit from such personality profiling, which could better link them to the right industries and jobs in those sectors. In simple words, computers could streamline job recruitment, and role mapping, and even critical things like interpreting large buckets of data to come out with interesting anecdotes about human behaviour (in a far more qualitative way) than surveys etc., ever could!

Let’s hope this technology can be harnessed and put to better use in the years to come! But, sensibly!


Originally published at Chip-Monks.