Face Recognition and Facebook's Recurring Privacy Problem
Once again, Facebook has messed with users' privacy in the name of a new feature.
The latest controversy is over Facebook facial recognition, which can automatically tag friends in photos just by matching the image to a massive database of faces.
Face recognition is a useful, time-saving feature -- at least when it works. But it's also a creepy addition to Facebook that opts you in automatically. As my colleague Sarah Jacobsson Purewal reported, you can only opt out of getting automatically tagged by friends. The database can still technically match your name to your face.
Therein lies Facebook's big dilemma, the one that comes up time after time, with each new change to the site that demands more of users' personal information: Yes, letting users opt-in to new features would be a more respectful approach. But because Facebook is inherently social -- that is, it relies on the participation of many users -- opt-in is much trickier to pull off. In some cases, it's just impractical.
Take, for example, the "instant personalization" feature introduced last year. This allows partnering Websites to use and display information from your public Facebook profile, and from your friends' public profiles. For example, if you write user reviews on Rotten Tomatoes or Yelp, your friends can see those reviews when they visit the site, provided they're logged into Facebook. Had Facebook made this feature opt-in instead of opt-out, most people wouldn't have bothered. That would defeat the purpose of personalization, which relies on having lots of recommendations from people you know.
A simpler example is Facebook's broader attitude toward public vs. private information. In late 2009, Facebook made changes to its privacy settings to put an emphasis on "everyone," so that users would share their status updates with the entire Internet by default. In making this change, Facebook was trying to be more like Twitter -- a massive, ongoing, public conversation between lots of people, regardless of whether they're friends or strangers. I like Twitter, and I understand by Facebook would want to make this change. But again, it only works if a critical mass of people are participating. That's why the "Everyone" option for status updates is opt-out, rather than opt-in.
With facial recognition, Facebook faces the same dilemma. Facebook could give people the choice to opt in to its photo recognition database, but then how many people would bother? The whole point of Facebook facial recognition is to tag all of your friends in a photo without any manual work. If most of your friends aren't participating, the feature is worthless.
I'm not defending Facebook's actions, but I understand why the site behaves the way it does. As long as Facebook introduces new features, there will be new privacy snafus. Facial recognition wasn't the first, and won't be the last.