The Risks To the Privacy of the 'Like' on Facebook
Political preferences, personal tastes, sexual orientation, religious beliefs, personality traits, emotional stability, ethnicity, age, sex or IQ,..., although not aware of them, if you have an account on Facebook and often click regularly on the icon 'I like it' in newspaper articles, events, products, celebrities, pictures or comments from friends published on this most famous social network, that will provide a lot of information which is recorded and analyzed as a whole, and can faithfully reconstruct a portrait of you. And that can be a threat to your privacy.
A study of 58,000 users of Facebook in the US has warned that the techniques available today to analyze the information we provide can not only be used to deliver advertising contents, but also steal potentially sensitive information about any users from Internet.
The study, published in the journal Proceedings of the National Academy of Sciences' (PNAS), is signed by researchers at the University of Cambridge (UK) in collaboration with Microsoft Research Cambridge .
As the authors note, almost without be aware of it, users who regularly click icon 'Like' voluntarily disclose information that help develop a fairly reliable profile of their personality and tastes. Users also indirectly reveal other personal data, such as whether they are consumers of addictive substances or their parents are separated.
The researchers developed statistical models that determine personal information about users through the "Like". Moreover, these results were compared with demographic data and questionnaires carried out in an application available on Facebook called 'My Personality', which allows the users to compare their personality traits with friends after responding to various issues.
When determining whether it was male or female, the coincidence was 88%. In 82% of cases, the system hit when the user catalog as Christian or Muslim, whereas classification by marital status or substance abuse was a coincidence of between 65 and 73%, according to the researchers.
In 95% of cases the system properly distinguished African-American individuals from those of Caucasian origin.
The authors also stressed that not many users when they click on icons, it explicitly reflects this information. For example, only 5% of gay people who took part in the study clicked on the icon to show they liked gay marriage.
The researchers also measured the correspondence between the subjective personality traits and therefore it is more difficult to evaluate, such as intelligence, emotional stability, or if users were open and outgoing.
His conclusion is that this type of research can be a revolution for the study of psychology as it allows obtaining data of very large samples of people worldwide at a cost much lower than traditional methods of research centers.
However, Michal Kosinski, one of the authors of the study, warned that, despite the advantages of using new technologies, all this personal information provided through social networks poses risks to privacy. On the other hand, it allows similar predictions from other digital data, so that it could be used to obtain personal information about citizens who do not want this information to be revealed.
"I'm a big fan and an active user of great new technologies, including Facebook. I appreciate the automatic book recommendations, or let Facebook select the most relevant in my own stories. However, I can imagine situations in which the same data and the same technology used to predict the political preferences or sexual orientation, be assumed to be a threat to the liberty or even life" says Kosinski.
Therefore, this researcher from the University of Cambridge claims that social networks offer users "the control over their information". He hoped that the results of this research would remind Internet users to be wary of displaying personal information online .Therefore, it encouraged "to use privacy controls and never share your content with users who you do not know".
Follow Charles Crawford on Twitter