One day you may wake up and there will be thousands of photos of you, which you didn’t know existed, available to anyone with just a few keystrokes.
Now, what does that possibility mean for you?
I started writing this piece 5 years ago in 2008, when I’d been thinking about the wealth of photography flooding the web, as well as the emerging facial recognition technology. Returning to the subject now in 2013, these thoughts seem at once less novel, but all the more relevant: As we immerse ourselves more deeply online, as we share ourselves more willingly, as technology evolves, which allows us both to create content anywhere we go, at any time, and to recognize patterns (and specifically people) in data, we’re approaching some sort of tipping point in which much about ourselves, which we thought private is suddenly going to become public. We may be approaching a point where content about us we didn’t even know existed is suddenly, easily available on the web. Suddenly, we will be ubiquitous.
What do I mean?
We’ve been enjoying (or enduring depending on your perspective) a continuing explosion in the social networking arena for a while now. But if you think social networking sites like Facebook and Twitter have brought extraordinary scrutiny to your personal details in exchange for an efficient means to communicate with your life’s collection of friends, business associates, acquaintances and hangers-on, well, you ain’t seen nothing yet.
In this period when digital cameras and mobile devices have proliferated, it’s not unusual to see folks snapping away at any lunch, brunch, dinner, party, or intersection of bodies coming through the rye. But have you ever considered that often when it’s not you or your friends taking the photos that you’re still making cameos in other people’s photos? Then what happens to those photos? We don’t know, so we usually don’t give it much more thought than that. Probably they end up on someone’s computer, maybe get circulated among a handful of friends. We shrug. Maybe they even get uploaded to Flickr or Twitter or Facebook, but who cares, right? Because those people don’t know your name, so they can’t tag you in the photo, so no one will ever find you.
What does this mean?
Well, it might mean that there are dozens of photos of you on the Internet you don’t know about. If you really get around, it may mean hundreds. Thousands? Who knows. (No, literally. Who knows?) If you live in a city like New York and walk through Times Square regularly, how many tourist photos do you think you end up appearing in ?
Your mug might be in [who knows how many places] on the Web and you wouldn't even know it. And to a certain extent, who cares right? No one’s like to ever stumble across those photos and if they did, no big deal. Good for a laugh.
But what if there were a systematic way to search photos and to recognize and label people within them? That really doesn't sound far-fetched to you anymore, does it? (It may have 5 years ago, when I first started this piece.) After all, Facebook and Instagram recognize where faces appear in photos now and ask if you want to label them. Apple’s iPhoto recognizes and suggests tags for specific people. Google’s Picasa app started doing this in 2008.
What seemed a bit like something out a Bourne movie a handful of years ago is becoming increasingly pedestrian. Of course, this technology has existed since the 1960s and has become increasingly more refined ever since. It was used somewhat controversially in the 2001 Superbowl (consider that this was pre-9/11) to identify potential criminals and terrorists, apparently resulting in 19 arrests. It’s been coupled to the CCTV system in the London Borough of Newham since 1998, despite criticism of its efficacy. It was used after the recent Boston Marathon bombings with little success. However, this study argues that law enforcement should have taken greater advantage of facial recognition software to more quickly identify the bombers, specifically Dzhokhar Tsarnaev. New technology allows faces to be recognized even with a few pixels. So you don’t even need a clear photo to identify people.
The technology is based on simple pattern recognition. Our brains do it for us: That’s how you recognize your friend when he’s headed down the sidewalk or your lover’s face lying beside you in bed or Barack Obama behind a podium on TV. And computers are learning to do the same thing awfully well. A facial recognition program can analyze some 80 points on your face, as well as the distances between them to arrive at what can be a highly accurate match. The result is a “faceprint.” And facial recognition software has evolved to the point that it’s something you can download to your computer. Now, imagine a point not too far down the road: In addition to Instagram, you can install a facial recognition app. Or imagine Google image search with facial recognition. Then you simply enter someone’s name or a photo of them or both and it trolls the Internet for any photo or video including their appearance across the entire Internet. If you give it a name, it matches that name against any existing photos of that person it can access on the Web. But you don’t need a name. You just upload a photo from your desktop and it starts spitting back photos of that person, including many, many photos that person didn't even know existed.
Now, some of us are saying — perhaps especially our more exhibitionist Gen Y friends — “No biggie, I got nothing to hide.” In fact, in response to the recent Snowden leak about the NSA’s activities in this country, many are saying exactly that. Someone created an account to cover such mentions on Twitter, of course. “Nothing to hide”? Well, good for you. I’d probably, generally fit into that category myself. But what about the closeted gay person, who can now be seen dancing in a gay club? Someone whose family may not speak to them again after finding such photos. What about the tourist traveling abroad who drinks a beer or visits a strip club only to return to a country, whose religious leaders punish people for such activities? What about the woman caught on camera stepping into an abortion clinic? Hey, what about the politician or priest leaving some place he or she would rather not be seen leaving? Now they can be named and perhaps their paths even traced.
If this technology blossoms into the user-friendly mobile or desktop app, which it almost certainly will, its use will be an extraordinary step forward for a surveillance society. (Don’t forget to introduce Google Glass into the equation, either.) So far we've been willing to make trade-offs. We allow acquaintances to see some personal things about us on Facebook — folks we may have little in common with and with whom we’d normally not share much information — in exchange for quick, economical, efficient, cheap communication across a wide network through a single channel. But there’s a lot of other information — and images — about you and me out there that simply hasn't been tagged or recognized yet. Soon — any day, any minute maybe — the opportunity will exist for all of that content to be found and tagged.
So, one morning you may wake up to find the Internet flooded with photos of you, which you didn’t even know existed.
Consider that this effect would be retroactive: Images could be surfaced of you, from the past, which you were never aware of. If you took steps right now to somehow ensure your photo were never taken again, there may still be hundreds, thousands of images of you extant on the web. You’re virtually (digitally) ubiquitous, whether you like it or not. At a concert. In the park. At a restaurant. On the sidewalk. At the beach. Out shopping. On vacation. In your church. Got your Facebook locked down? Got an avatar instead of an ID photo on your Twitter profile? Doesn’t matter. All that’s needed is one image of you on the Internet coupled to your name and every image on the Internet can be coupled to you. And if you’ve so much as stepped outside in the last 10 years, your image is almost certainly on the Internet.
Something else important to remember then: I probably won’t need your name to uncover your identity: If I have your image, I can find your name. In an excellent his presentation “Privacy in the Age of Augmented Reality” at SXSW this March, Carnegie Mellon professor Alessandro Acquisti showed how with a single photo he could track down, not only someone’s name via facial recognition (all he needed was one tagged image on Facebook), but given an image and a name, he could quite quickly get most of a social security number. And often without much work, he could find more.
Is it all coming clear? Perhaps, we as a culture need to have a conversation about where all of this perfect storm, this collision of social media, digital photography and facial recognition is headed. Because the privacy implications are monumental.
Now for the really portentous stuff. I ask this somewhat seriously. Are we on the verge unwittingly replacing “God” with our own secular, collective observer in the form of incessant digital monitoring?* An Orwellian “god device” which trains an unblinking eye upon us? Potentially, keeping us in line? Potentially, smacking us for our non-conformity or for our misdeeds? (For something as simple picking your nose in public. For something as integral to your identity as being gay. For thinking you were incognito when acting against expectations in public. Choose your poison.) Well, this scenario would be Orwellian, except for one important difference: We've not been forced to surrender to some sort of fascist government’s rule a la 1984 in order to enable this ever-watching eye. No, we've willingly, gradually submitted to it our selves. Transactionally. We've traded our privacy for dribs and drabs of entertaining data, for the ability to share ourselves. Can we really complain if there’s a little blowback coming our way?
*Jeremy Benthem’s panopticon comes to mind also, of course.
Call It Super Bowl Face Scan I — Wired, 02/02/01
Revamped Google Picasa site identifies photo faces — CNET, 02/09/08
Engineers Test Highly Accurate Face Recognition — Wired, 03/24/08
More Than Facial Recognition — Carnegie Mellon University, Summer 2011
“Hallucinating” a face, new software could have ID’d Boston bomber — Ars Technica, 05/29/13
A Case Study on Unconstrained Facial Recognition Using the Boston Marathon Bombings Suspects [PDF] — Joshua C. Klontz & Anil K. Jain, Michigan State University, 05/30/13
Google Irks Developers with Ruling on Facial-Recognition Apps — MIT Technology Review, 06/10/13
Upcoming Biometric System in Helsinki: Pay With Your Face — Core77, 07/22/13
You can listen to a podcast of my discussion on this subject with University of Wisconsin — Madison students. Recorded July 24, 2013.