Worried About Privacy? Staying Off Facebook Won't Help

It's hard to feel safe on Facebook sometimes. For most people, social networking is the easiest way to keep up with friends and family and worrying about fake offers, spam, and apps harvesting personal data just get in the way. However, recent research suggests that for the privacy-obsessed, not joining Facebook and similar sites is no longer the answer.

A quartet of German researchers was able to learn about people who weren't members of social networks by examining information posted online by their friends, according to a recent paper published in the journal PLoS One. Researchers Emöke-Ágnes Horvát, Michael Hanselmann, Fred A. Hamprecht and Katharina A. Zweig looked at publicly available information on Facebook to determine with an "astonishing" rate of accuracy whether two non-members knew each other based on information posted by their friends.

People worried about revealing details about their lives, political views, or other sensitive information online often stay away from social network platforms "in the belief that will help protect their privacy," the researchers wrote in the paper's introduction.

"Such an assumption is no longer valid," they concluded.

With the help of machine learning, social network operators can make predictions about individuals being mentioned by users, even when the person in question was not a member, the researchers said. For the purpose of the study, researchers used "real-world Facebook friendship networks" representing students from five universities in the United States.

Friends Post About You
Even if people don't post pictures of themselves attending a party on Facebook, their friends might. While this study focused on guessing whether or not people knew each other, it is not a far reach to imagine different ways friends' data can be used to create a profile of non-members. It's a fact that many users are unaware of how much of their information is public and being shared, so it's possible that incriminating photos are out there for anyone to find. Or friends checking into a location and mentioning who they are with. The Wall posts are a goldmine of personal information.

A Carnegie Mellon researcher demonstrated at last year's Black Hat conference how he was able to apply off-the-shelf facial-recognition software to Facebook photos to identify 30 percent of students walking around CMU campus. Alessandro Acquisiti took his research another step further and correlated the publicly available information with anonymous profiles on online dating sites. He was also able to correctly guess Social Security numbers within four tries for 28 percent of the subjects. And all this started with just a single photo.

German data-protection officials have even requested Facebook to disable its facial-recognition software.

It's a little funny to me that joining Facebook may actually help with protecting user privacy. Facebook allows people to tag their friends in photos, and as of yet, there is no way to opt-out of getting tagged. Right now, I have Profile Review enabled in Facebook so that I am notified whenever someone tags me in a photo. I manually remove the tags because I know many of my friends have fairly public profiles. I have location check-in disabled, and last I checked, that meant someone else couldn't check me in, either. If I wasn't on Facebook, I wouldn't have any control over what's going on about me.

Security Questions
Consider how many security questions for online services rely on personal information that people casually toss out online. Now imagine if that can be linked to someone else. My colleague Neil Rubenking wrote recently about how kids have a digital presence before they are two. It's one thing to play fast and loose with your privacy, but it's totally a different issue when it's someone else's.

Play fair and take advantage of those privacy settings on your social networks. Stop and consider who is actually going to see that information, because one it's out on the Internet, it's out there.