The results of a Cambridge University study show that a computer can determine people's personalities by analysing the Facebook posts that users like. Researchers have consequently been able to invent an algorithm that uses Facebook interactions to predict personalities—and the technology has proven to be accurate.
The researchers used openness, conscientiousness, extroversion, agreeableness, and neuroticism—known as the "big five" psychological traits—to develop the software. The team first compared the predictive abilities of the software to those of work colleagues, and found that the former was more accurate after looking at 10 'Likes' on the social media platform.
The team then increased the number of 'Likes' that the software analysed. The researchers found that 70 'Likes' resulted in a more accurate prediction of character than that of a friend, while 150 'Likes' performed better than a parent, sibling or partner. For the software to supersede the capabilities of a spouse, 300 'Likes' were required.
The researchers recruited 86,220 volunteers on Facebook, who were asked to complete a personality questionnaire that consisted of 100 items. The volunteers also provided the researchers with access to their 'Likes'.
Lead author Wu Youyou said that the study indicates that "the emergence of emotionally-intelligent and socially skilled machines" is a possibility in the future. Ms Youyou identified job recruitment as one area in which benefits could be seen, as recruiters "could better match candidates with jobs based on their personality."
The results of the research were published in the 'Proceedings of the National Academy of Sciences'.