Facebook Rolling Out Facial Recognition With Opt Out Privacy Setting That Could Irk Governments

In December Facebook launched a facial recognition feature for North American users that suggests friends to tag in photos. Facebook is now apparently executing a wider roll out of the feature with an opt out privacy setting that raises some privacy concerns.

The feature will likely increase the number of people tagged in photos, generating more compelling content and interconnections on the site. However, it could also draw government intervention from countries sensitive to privacy issues such as Germany which requested changes to Facebook’s Friend Finder, and Switzerland which is threatening to block Google Street View if it doesn’t obscure faces and license plates.

When those who’ve received the facial recognition feature upload photos, Facebook’s algorithms groups the photos by those with similar faces. It also suggests friends those faces may belong to by matching them with previously tagged photos of friends, and the suggested tags are automatically applied unless removed. This speeds and simplifies the tagging process by reducing the frequency with which users are forced to type out names.

Without facial recognition, photo tagging can be a chore, preventing users from finishing a task that produces notifications that bring friends back to the site as well as compelling news feed content. Considering well over 100 million photos are uploaded to Facebook a day, streamlining photo tagging presents an enormous opportunity for Facebook to increase engagement on the site and improve the user experience.

This may be why Facebook is pushing facial recognition, a concept with negative privacy stigmas, as an opt out feature called “Suggest photos of me to friends”. If the same functionality was opt in and was labeled “facial recognition photo tagging” or even included the words “facial recognition” anywhere in the 100-word description of the feature, few users might agree to enable it.

Facebook explained that the feature was opt out in its blog post announcing the feature, but that was six months ago. In the meantime, those outside the initial roll out could not preemptively opt out, and the recent extension of the roll out and the feature’s new inclusion in the privacy settings of hundreds of millions of users came silently. Most users can now visit Privacy Settings -> Customize settings -> “Suggest photos of me to friends” and then click edit settings to view information on the feature and toggle it between the default Enabled and the optional Disabled.

This quiet push combined with the inherent concerns around facial recognition of the world’s largest collection of photos might trigger complaints from some governments, particularly in Europe where there’s been closer scrutiny of Facebook’s privacy practices. The governments might request more transparency around the fact that the feature uses facial recognition, that it default to opt out, or a disabling of the feature for their residents in the most extreme case.

The feature’s ubiquity due to its opt out nature could benefit users looking to manage their reputation online, though. Users often omit tags of friends in scandalous or objectionable photos out of respect, but this can lead those depicted in the photos to never even know they exist. This prevents them from asking the owner of the photo to delete, leaving racy shots up for others to see.

Facial recognition being opt out is a prime example of Facebook’s internal philosophy of pushing privacy boundaries for the good of the user experience coming into conflict with more cautious, conservative beliefs. Soon the world might see suggested photo tags as a crucial part of Facebook, the same way the initially contested news feed is now. However, Facebook needs to accompany roll outs of opt out features with plenty policies of announcements and education if it wants to be seen as trustworthy.