Six Facebook Privacy Blunders
Facebook routinely pushes privacy boundaries and riles privacy advocates.
When Facebook introduced its auto-tagging facial recognition feature, users and privacy advocates around the world were immediately perturbed - not by the feature itself necessarily, but that Facebook rolled it out without telling users it was coming. EU privacy regulators also wanted to ensure that pictures were tagged only with the prior consent of their subjects and not automatically. Otherwise, the feature ran afoul of the law, they said.
But the facial recognition technology's strength or weakness is less of an issue, it seems. According to a recent study from Carnegie-Mellon University, the problem lies in how easy it is to re-identify someone using a snapshot and publicly available data.
The Wall Street Journal reported last week that Prof. Alessandro Acquisti used snapshots of students taken with a web cam, facial recognition technology similar to that recently acquired by Google, and publicly available Facebook photos to correctly identify nearly one-third of the student volunteers who participated in the study. Going further, he also demonstrated that with facial-recognition technology and publicly available information (such as information in a subject's Facebook profile), he could correctly predict the first five digits of the subjects' Social Security numbers 27 percent of the time.
This paper really establishes that re-identification is much easier than experts think it's going to be.
What have the privacy experts said about "anonymized" data all along? It's never really anonymous. With the right combination of "anonymous" information, anyone can be identified.