Computers 'know you better' than friends

Computers can be a better judge of character than our family and close friends by analysing Facebook 'likes', new research has suggested.  

The new study, published in the journal PNAS, used Facebook 'likes' as input data to predict five key character traits better than friends, parents and in some cases even partners.

According to researchers from the University of Cambridge and Stanford University, the findings are an “important milestone” in the social human-computer interaction. By analysing pure data, machines can know us better than formerly believed.  

"In the future, computers could be able to infer our psychological traits and react accordingly, leading to the emergence of emotionally-intelligent and socially skilled machines," said Wu Youyou, lead author from Cambridge's Psychometrics Centre.

The experiment involved 86,220 volunteers on Facebook who completed a 100-item personality questionnaire as well as providing access to their 'likes'. It was estimated that a Facebook user has an average of 227 'likes', a good enough kind of artificial intelligence (AI) to predict the personality accurately.

"The ability to judge personality is an essential component of social living – from day-to-day decisions to long-term plans such as whom to marry, trust, hire or elect as President," said Cambridge co-author Dr David Stillwell.

"The results of such data analysis can be very useful in aiding people when making decisions."

The users then had the option of inviting friends, siblings, parents or partners to take a shorter version of the personality test to assess their psychological traits.

The results’ analysis was based on five characteristics known as the OCEAN model in psychology: openness, conscientiousness, extraversion, agreeableness and neuroticism.  

By examining 10 'likes', the computer was a better judge of personality than a work colleague. Looking at 70 'likes', the machine managed to rival a friend or housemate, and it needed 150 to compete with a family member. Partners proved to be the most difficult to outperform with 300 'likes'.

"Big Data and machine-learning provide accuracy that the human mind has a hard time achieving, as humans tend to give too much weight to one or two examples, or lapse into non-rational ways of thinking," said Dr Michal Kosinski, co-author and researcher at Stanford.

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them