Facebook Shuts Down Massive Photo Tagging Facial Recognition Program

“We’re shutting down the Face Recognition system on Facebook," which has the faces of more than 1 billion users.
Image: Kevin Dietsch/Getty Images

Facebook is shutting down the photo tagging facial recognition program that was key to its product for years, and is deleting more than a billion “face prints” of users, according to a blog the company posted Tuesday. 

“We’re shutting down the Face Recognition system on Facebook,” said Jerome Pesenti, vp of artificial intelligence at Facebook. “As part of this change, people who have opted in to our Face Recognition setting will no longer be automatically recognized in photos and videos, and we will delete the facial recognition template used to identify them.” 


Specifically, Pesenti is referring to Facebook’s photo tagging system, which automatically identified a user and their friends in photos they uploaded to the platform and suggested that they be “tagged” in the photos.  

For years, photo tagging was a core Facebook feature and was one of the main things people “did” on Facebook. To tag a photo, you clicked a specific part of it and wrote in the name of your friend. This then added the photo to your friend’s page of tagged photos, and allowed them and their friends to comment on them, cementing Facebook as one of the main places people uploaded photos online.  

While photo tagging was manual at first, in 2010, Facebook began automatically detecting who was in a photo by comparing new photos to photos that were already uploaded of that person. This facial recognition technology made it much easier to tag photos, it was also seen as very creepy by a lot of users and civil liberty experts as Facebook automatically opted users into the program and used existing photos of them to build a database of face recognition templates.  

Face recognition became so embedded in Facebook’s design that it became a core security feature. If a user tried to log into Facebook from an unrecognized computer, they would be presented with random tagged photos of their friends and would have to correctly identify who was in the photo. It also eventually began notifying people who were in photos that they weren't tagged in. 

The move away from facial recognition in this specific instance, following a growing movement against the technology, is laudable, and represents the end of what is one of the largest face recognition systems ever created, considering Facebook’s more than 2.9 billion users. But Facebook said it “still see[s] facial recognition technology as a powerful tool” for preventing fraud and impersonation, for example, and as Motherboard wrote Monday, the company is still trying to monetize basic human activity through its metaverse plans. Facebook also told the New York Times that it is not getting rid of DeepFace, the software that powers its facial recognition system.

What’s also important to recognize is that Facebook’s core product has long de-emphasized the mass uploading of photos and has moved toward sharing of news and statuses that can be endlessly bickered about on the news feed. Facebook has become so bloated that it can be difficult to even find your own photos, and Instagram is now Facebook’s photo-specific app. So what we have here is a move that comes only after Facebook spent more than a decade mining the faces of a large percentage of the humans on planet Earth to help its own growth and power, and only comes after the platform itself has evolved past the need for this type of technology.