As this practice becomes more ubiquitous, many privacy advocates are raising concerns. “Face recognition is every bit as sensitive as geolocation data, or logs of whom you've called,” said Alvaro Bedoya, founding executive director of Georgetown Law's Center on Privacy and Technology. And while you may be able to leave your phone at home, he said, you can never delete your face.
Facebook has already been scanning faces in photos to suggest which of your friends you should tag in pictures. But the new feature will recognize you without any assistance, although it will only work if you've already agreed to the tagging feature.
Facebook's database of facial data has an impression of people's faces, rather than actual images, Joaquin Quiñonero Candela, the company's director of applied machine learning, said in a blog post. Any photos or videos people upload to the site are then matched against that impression.
Now, all of the Facebook facial scanning features are under one privacy setting. That means that if you want to receive notifications about where your face appears on users' accounts, you'll also have to be fine with Facebook suggesting to your friends that you be tagged in photos.
That change drew criticism. “Facebook tried to simplify the option to make an on-off feature but doesn’t allow users to choose what features they may want to use,” said Jeramie Scott, national security counsel at the Electronic Privacy Information Center. Facebook deputy chief privacy officer Rob Sherman acknowledged in a separate blog post that the company's “all-or-nothing” approach may not sit well with everyone, but he said the company believes it's better for people “to manage one master setting.”
Besides notifying individuals when they appear in photos posted by friends, Facebook will also tell you who — friend or not — uses a photo of you in their profile picture. This is a move to curb impersonations, which are already prohibited but can be difficult to detect. Finally, the company is also going to use facial recognition to describe photos to the visually impaired; someone looking at a photo using a screen-reader can hear which friends are in the picture by name.
Bedoya said he thinks some of the new features offer good fixes, just to a bad program. “While these individual changes are positive in that they make things clearer for users, they have to be set against Facebook's widespread flouting of industry standards on facial recognition,” he said.
Facebook apparently does not see a clear standard on how companies roll out these features. “We recognize that there are a lot of different views on this issue and lots of different approaches to providing people with transparency and control,” the company said in a statement. “We decided, in consultation with privacy advocates and people, that the best approach for this offering was to notify people on Facebook and provide a way to adjust it in their account settings at any time.”
The social network has been sued by people who claim that the face-scanning technology used by Facebook violates state laws that say companies may not keep biometric data — face scans, fingerprints, etc. — without first getting permission. A California judge last year allowed a case from three men in Illinois to move forward over Facebook's objections. And the newest features will not roll out in Canada and Europe, where regulators have raised concerns about the way Facebook collects face data.
The features, which began rolling out Tuesday in the United States, will apply only to new photos uploaded to the site and will not scan through existing photos. Facebook will let people know about the changes in their news feeds, the company said, and will encourage everyone to check their settings.