Picture recognition, social networks, mistakes, and privacy-enhancing technology

One of the most fascinating side conversations last week was about picture and face recognition in social networks: The basic argument was that, even if people might have the good sense to use a pseudonym for some social network activity, they’ll often associate photos with their profiles, and that these photos might lend themselves to easy recognition and linking of online profiles to real-life identities, undoing the privacy effect of having chosen a pseudonym in the first place. It might be a good idea, that particular argument continues, to include features with social networking sites that might obfuscate photos to the point that they aren’t easily recognized by machines. As an example, a somewhat embarrassing questionnaire scraped off some social networking site was shown. (“Have you been involved in a fist-fight? Have you been arrested? …”) It featured a photo of what was presumably the young lady that had filled it in. The implication was that, some time soon, a would-be employer might find that entry based on a photo search, with all that follows.The “let’s obfuscate photos” approach is a particularly illustrative example for a deeper conflict: The one between self-expression online (which is enabled by sharing real data, real photos, real stories — and real lies!), and privacy online (which is endangered by sharing real data, real photos, real stories — and real lies!). The question in this conflict is no longer about collecting (or publishing) data (as the obfuscators seem to suggest). The solution is no longer about pseudonyms or hard controls: The question is ultimately how we deal with the societal and personal cost that comes with the Net’s vast opportunities for self-expression, and the solution must be in the same sphere. Technology can help us find a solution, though, and Creative Commons is an important example for this approach in the copyright space: It helps people to do the right thing. And sometimes it helps people to make errors that need to be settled in court.The critical observation here is that technology takes an helping role, not an enforcement role, not a decision-making role. The decisions and the enforcement are kept on the social and societal level.And to some extent, we might be starting to see societal change: When youthful (or later) mistakes are made part of a permanent record by the information sphere that we live in, when 20 year old traffic tickets show up in the reports that would-be employers ask to be prepared about candidates, we might find that the answer is actually to just cope with these mistakes, and to accept that to be human means to err.