Stay away from social networks and people won't know who you're hanging out with or what you're doing, right? Wrong. When it comes to social networking, a recent study suggests, you can run but you can't hide.
A paper published last month in the journal PLoS One shows how researchers were able to learn about nonmembers of social networks based on information their friends posted online. Using machine-learning models, German researchers Emöke-Ágnes Horvát, Michael Hanselmann, Fred A. Hamprecht and Katharina A. Zweig were able to predict whether two nonmembers of a social network knew each other based on information shared by a mutual contact on the network.
In other words, even if you’re one of those holdouts who refuses to join Facebook and other social networks due to privacy concerns, the data your friends share is enough to let anyone with access to that data draw conclusions about you. And while the initial research in the area focuses on the relatively innocuous facts surrounding who you do and don't know, it will become increasingly easier to draw profiles of people based on what their contacts share on Facebook.
“To our knowledge these are the first results on the potential of social network platforms to infer relationships between non-members,” the researchers wrote. They also noted that the relationships were predicted with an “astonishing” rate of accuracy simply by scanning readily available information on Facebook for students at five U.S. universities.
And the authors were working only with publicly available information. Social networks may have a vast trove of data about members that isn't generally available. “Social network platform operators typically have access to much more detailed information on nodes such as the age, sex and (approximate) location of their members; and if they provide messaging services they can infer the quality of an acquaintance from its communication pattern,” they wrote.
Researchers have long known that studying real-world social networks is a good way to predict individual behavior. We've previously reported on how online social networks can be used to predict a person’s risk of contracting a sexually transmitted disease. But the latest findings suggest a path toward an all-encompassing model that may one day be able to predict much more than who you know.
“Ultimately,” the study concludes, “it evokes the question of the ownership and exploitation of relational data in the information age.”
Photo courtesy of Shutterstock.