Facebook might have finally figured out a way to wipe out the annoying photo-tag spam that plagues the social network.
Everyone has seen it: An advert for knock-off sunglasses posted in a Facebook group that tags you and a hundred other people. One of your friend’s accounts, apparently hijacked, posts a scam and tags all their contacts. A “tag yourself!” infographic of little-to-no interest to you. And so on.
Whatever its form, it’s ugly and annoying, clogging up your news feed and your notifications. And Facebook is trying to stamp it out.
It does this by analysing the photos themselves for clues as to whether the tags are “legitimate.” The system can look at whether the photo was captured “natively” or uploaded from a computer, or the social relationships between the tagged users — lots of users with nothing in common with each other and no known interactions could be a red flag for spam, for example. It can also look at whether (the correct) faces are in the photo where they’ve been tagged — because a spam advert tagged with lots of people might not include any.
So if detects an “illegitimate” photo, what then? The system can take a couple of options.
It could take the nuclear option and not allow it to be uploaded in the first place. Or, it might allow the upload, then “suppress” the image so it does not appear on newsfeeds of other users or the groups where it was uploaded. Another alternative is to just block the notifications — so even if you’re tagged, it doesn’t clog up your Facebook experience with spam.
Whatever the approach, it should result in a cleaner, less annoying Facebook experience.
It’s worth bearing in mind that just because Facebook has applied for a patent in this area doesn’t mean this feature will find its way into a finished product. Companies apply for thousands of patents every year that never get implemented — they might be filed as a precaution, or to trip competitors up in litigation.
But spam is a serious problem for Facebook, and this isn’t the first time social network has thought about ways to fight it. A previous patent application published in November 2016 details how Facebook could crack down on fake goods being sold on its platform by scanning images for brands. If a Facebook page was using the Ray-Ban logo without permission (even if they were cleverly not using the brand name in text) to sell sunglasses, the system could flag it up and take it down.