People's intimate personal details which they may not wish to share publicly – such as their gender, sexual orientation and drug use – can be reliably guessed at, by analysing their Facebook likes, according to new research.
A team from the University of Cambridge's Pschometrics Centre, working with colleagues from Microsoft Research, analysed 'like' patterns of 58,000 US Facebook users, who volunteered to share their data.
Their models were able to predict male sexuality with 88 percent accuracy, identify political affiliations with 85 percent accuracy and even spot whether a subject was likely to use recreational drugs with between 65 percent and 73 percent accuracy.
Tellingly, the researchers were able to make these predictions even though the vast majority of likes users made were not for explicitly revealing links: less than five percent of gay users reviewed liked topics such as 'gay marriage'.
The vast majority of likes were for music, TV shows and films.
The implications are not just linked to Facebook, said Michal Kosinski, operations director at the Psychometric Centre, with people potentially leaving a vast digital crumb trail across numerous sites, that could be used to discern revealing information.
"Similar predictions could be made from all manner of digital data, with this kind of secondary 'inference' made with remarkable accuracy - statistically predicting sensitive information people might not want revealed,” he said.
“Given the variety of digital traces people leave behind, it's becoming increasingly difficult for individuals to control.”
The results threw up some odd links. For example, those that liked curly fries were more likely to have a high IQ; those that clicked 'like' on a link entitled 'That spider is more scared than u' were more likely to be non-smokers.
Kicking Palantir off of AWS is among their demands, too
Rafaela Vasquez was watching The Voice at the time of the crash, new evidence shows
PUBG price slashed on Steam after selling more than 50 million copies - as daily player numbers plunge
Use the same password for every website? It might be time to change them all