Getty

Does Hiding Your Face With Your Hand Actually Hide You From Facebook?

A facial recognition expert on the ‘concerning’ future of facial analysis tech, and the best way to keep prying eyes from finding you on social media

Whether it’s Kylie Jenner’s iconic Snapchats or the art of the dab, teenagers covering their faces in photos is not anything new. Recently, a new batch of photos surfaced of British teens half-covering their faces with their hands in a nightclub. But why? Are they trying to hide spots of acne? Is the flash of the camera too bright? Or is this just the Sexy New Thing To Do? A deeper investigation (a.k.a. googling) suggests that while some believe this particular ~look~ is a new trend, others would argue that the action of covering your face with your hand is a not-so-subtle, brilliant tactic for protecting one’s online identity, and more specifically, to avoid being tagged on social media via facial recognition technology. But does it actually work?

As it turns out, not really.

According to Jennifer Lynch, a lawyer and facial recognition expert at Electronic Frontier Foundation in San Francisco, the more you cover your face with your hands, the smarter you’re making a social platform’s algorithm. In other words, every time you upload a photo of yourself to Facebook covering your face with your hand, Facebook will take that information, improve its algorithm, and possibly use it to identify future photos of you doing the same thing.

But being tagged on social media doing a keg stand without wanting to be is just one example of growing privacy issues as facial recognition tech continues to advance. Lynch, who specializes in privacy and civil liberties issues in new technologies, explained how we’ll have to adapt as it improves, why having control over our images matters, and the one strategy that actually beats hiding your face with your hand.

Whether it’s just being silly or it’s for vain reasons, young people covering their faces in photos — especially in selfies — isn’t exactly a new trend. But in terms of covering your face specifically to avoid facial recognition on social media apps, is this something that is becoming increasingly common?

Jennifer Lynch: Well, I’ve actually never seen this before. So it’s interesting to me if kids, teenagers, or anybody is doing this to try to beat facial recognition.

Face recognition is designed specifically to look at the features on your face and surrounding the face and make an identification or verification based on those features. But if you are using something like Facebook to do face recognition, Facebook has this algorithm that does the identification but is constantly learning from its users’ feedback. So when Facebook presents an image to you and says, “Is this your friend Sheila?” you’re going to say yes or no and then based on that, Facebook is going to update and improve its algorithm. The reason that I mention that is because even though facial recognition is designed to recognize faces, if you are constantly uploading a picture of Sheila with a hand in front of her face, it’s quite possible that the face recognition algorithm could learn how to identify Sheila with her hand in front of her face. The algorithm doesn’t care whether it’s a hand or a nose, it’s just trying to match one thing against another thing.

Now, some social media will allow you to turn off face recognition. For example, Facebook will allow you to turn off their "tag suggest" feature. And, according to Facebook, if you turn it off you’re no longer using facial recognition on your face. So that might be a slightly easier way to get around it, but as to whether this is actually able to defeat it is a really interesting question.

So if you’re constantly covering your face with your hand, it’s almost like you’re defeating the purpose.

Lynch: Yes, and possibly even making it easier for the algorithm to identify other people with their hands in front of their faces. Because what the algorithm is doing is constantly learning.

Let’s say you have really long hair and then you chop it all off. The algorithm doesn’t know that that’s hair, but it’s looking for the shape of the hair and how it fits on your face. If you have enough times of uploading a picture identifying yourself with a shorter haircut, then it’s going to learn how to do that. And putting a hand in front of your face could very well be the same kind of thing. But the other thing that’s happening is that not only is it learning how to identify you better, but if this is in fact a trend of people putting their hands in front of their faces, the algorithm is going to start compensating for that and start to recognize that this thing in front of the face is something that is used to identify people going forward.

Some would argue that if you’re going to cover your face, why allow someone to take your photo in the first place? I’d argue that the point is that you’re in control of where your photo goes. Why is this type of ownership over your image so important?

Lynch: As a privacy scholar and advocate, the way that we tend to think about privacy is not that you have to keep everything secret, it’s not that you can’t share anything with other people — it’s that you have control, like you’re saying. You have control over the dissemination of the information about you, so that means that I might want to share a photograph of myself with 200 of my friends on Instagram, but I don’t want to share it with my employer. I don’t want to share it with the government. And that, really, is what privacy is all about — being able to control that. So I don’t think that people should have to refrain from posting images of themselves just to have privacy protection. It’s very difficult in today’s society to participate and communicate without using social media.

There’s a big misconception that teens don’t know how to protect their identities online, but the reality is they’re hyperaware. Having private Instagram accounts and using nicknames as Twitter handles is something teens have been doing for years, but as technology continues to advance and privacy is potentially reduced, how will young people have to adapt?

Lynch: I agree that it’s completely false to say that teens don’t care about privacy. It’s just that they see a different threat model. The threat to privacy might come from parents or another group of friends rather than from the government or an employer. So they’re going to be trying to develop privacy protective measures based on that threat model.

In terms of technology that teens could use going forward, I think that they’re going to have to become as tech savvy as the rest of us and learn how to use encryption and learn how to really adjust their privacy settings on social media to prevent people from getting into their accounts.

Where do you think facial analysis tech is heading? What’s on the horizon?

Lynch: What should be most concerning is that one of the things that face-recognition researchers are working on is how to identify people as they age. So that’s a huge thing for anybody who’s putting their image online now and allowing face recognition to map their image because it’s helping to develop the algorithms that will identify people over time.

Let’s say that a 15-year-old is the same person who is then 65. Your face changes so much over that amount of time. And really, it’s very difficult to even identify somebody if there’s 10 years' difference in a picture, especially if you’re talking about somebody who is 15 versus 25 or 5 versus 15. But that is definitely changing. Those technologies are improving and we will be able to identify people in the future based on their younger images. And you could think about that in terms of the things that we all do when we’re young that you might not want to hang around when we’re older and applying for jobs.

The other thing that’s concerning to me as somebody who works on privacy is that face recognition is being developed to identify people as they walk around in physical space. The government is building out big databases of people’s faces; some of those are coming from driver’s license photos. So any state that uses face recognition on its driver’s license database, those images could potentially be linked to security cameras and surveillance cameras. People could be identified as they walk around in public.

If you’re applying for a job, internship, or college, what’s the best way to cover your tracks?

Lynch: To any extent that you can lock down your own personal accounts to prevent people from accessing it that you don’t want to access it — that’s the first step that I would take. Deleting any photographs that you might think are embarrassing or might put you in a bad light. The other concern, though, is that your friends may have posted photographs of you that you’re not super happy with and that may involve going to your friends and asking them to take photographs down.

Seeing how quickly technology advances, do you anticipate an alternative — or improvement — that beats covering your face with your hand?

Lynch: Again, it’s a computer algorithm, so you could try to change your facial features or put your hands in different positions over your face, but ultimately, this is where it’s important to talk to and know your city government. Talk to state legislatures, talk to Congress about getting these things regulated. Because until there is a law in place to prevent the collection of this kind of data, the companies are going to continue to collect it.