Smile, You're On Everyone's Camera
Ubiquitious facial-recognition software is coming.
According to the Wall Street Journal, police departments across the nation will soon adopt handheld facial-recognition systems that will let them identify people with a snapshot. These new capabilities are made possible by BI2 Technologies, a Massachusetts company that has developed a small device that attaches to officers' iPhones. The police departments who spoke to the Journal said they plan to use the device only when officers suspect criminal activity and have no other way to identify a person—for instance, when they stop a driver who isn't carrying her license. Law enforcement officials also seemed wary about civil liberties concerns. Is snapping someone's photo from five feet away considered a search? Courts haven't decided the issue, but sheriffs who spoke to the paper say they plan to exercise caution.
Don't believe it. Soon, face recognition will be ubiquitous. While the police may promise to tread lightly, the technology is likely to become so good, so quickly that officers will find themselves reaching for their cameras in all kinds of situations. The police will still likely use traditional ID technologies like fingerprinting—or even iris scanning—as these are generally more accurate than face-scanning, but face-scanning has an obvious advantage over fingerprints: It works from far away. Bunch of guys loitering on the corner? Scantily clad woman hanging around that run-down motel? Two dudes who look like they're smoking a funny-looking cigarette? Why not snap them all just to make sure they're on the up-and-up?
Sure, this isn't a new worry. Early in 2001, police scanned the faces of people going to the Super Bowl, and officials rolled out the technology at Logan Airport in Boston after 9/11. Those efforts raised a stink, and the authorities decided to pull back. But society has changed profoundly in the last decade, and face recognition is now set to go mainstream. What's more, the police may be the least of your worries. In the coming years—if not months—we'll see a slew of apps that allow your friends and neighbors to snap your face and get your name and other information you've put online. This isn't a theoretical worry; the technology exists, now, to do this sort of thing crudely, and the only thing stopping companies from deploying it widely is a fear of public outcry. That fear won't last long. Face recognition for everyone is coming. Get used to it.
What's changed in the last decade? Three things. First, computers have gotten better at recognizing faces. The technology works by analyzing dozens of different features—the distance between your eyes, the width of your nose—that remain the same across photographs. As computers have gotten faster and digital photography has gotten better, face recognition has filtered down to consumer photo software. (I find Picasa's to be uncannily good). By this point, every tech giant has snapped up face-identifying expertise. Apple purchased the face-recognition firm Polar Rose last year. In 2006, Google acquired the biometric recognition company Neven Vision, and Hartmut Neven, one of the world's experts in computer vision, is a respected engineer at the company. A Microsoft research team in Israel has built a fantastic app that uses face-recognition systems to search the Web for pictures of people who are in your photo album. And last year Facebook rolled out a tool that automatically suggests names of people to tag in your pictures.
Another major factor that augurs the face-recognition era is that we've become accustomed to ubiquitous photography. Now that we all carry cameras everywhere, it no longer seems odd when someone points a lens in your direction—you probably don't even notice it. Indeed, we all have a general expectation that if we go outside, someone is going to capture our image. We might not like it, but we live with it.
Finally, there's Facebook. Ten years ago we were worried about authorities building a worldwide database of our faces. In 2004, a Harvard student built a site that let people post their own names and faces, and hundreds of millions of people around the world have taken him up on it. We're all posting pictures, and tagging names to pictures, at a furious rate—according to Facebook, people add 100 million names to faces on Facebook every day. The face-recognition tools available to law enforcement agencies will match you against government databases—the DMV or passport database, or the FBI's most-wanted list—but the technology available to consumers will be able to do just as well by matching your face to online snapshots. The government couldn't have built a better facial database if it tried. (See the Onion's video, "CIA's 'Facebook' Program Dramatically Cut Agency's Costs.")
So, when will consumer tech companies get over their fears of bringing out a mainstream face-recognition app? In May, Eric Schmidt, Google's executive chairman, told an audience at the D9 conference that the company had decided not to put out a face-recognition tool out of fear it would be misused. People at Google have indeed confirmed to me that the company has no current plans to release such a system. In fact, Google has gone out of its way to prevent face-recognition in its apps. Google Goggles—the smart-phone app that returns information from the Web when you take a photo of an everyday item—doesn't work on faces. But that's not because it can't work on faces. Google engineers have publicly admitted that, in prototypes, the system is really good at identifying faces. At a conference last year, Google engineer David Petrou said that the system is capable of identifying the face of someone who has as few as 17 other pictures online. * If there are 50 photos, Goggles will be able to get your name almost every time. "We do it well, but it's not deployed," Petrous said.
But Google won't have that luxury forever. Google reps have told me that face-recognition is something that Goggles users have asked for. This isn't surprising—such a tool would be incredibly useful. If you meet someone at a conference, you could point your camera their way and get their contact info. You could snap a group photo at your family reunion and have the pictures automatically filed by name in your photo album. Or you could think of it as Shazam for strangers—who is that woman who's always winking at you on the bus, or that creepy fellow who's leering at you from across the bar? Snap their pictures and find out. Given the demand, and given the widespread availability of the technology, it's inevitable that some company will release such a face-recognition tool very soon. And once the horse leaves the barn, it will become socially acceptable for Google and other tech giants to follow.
What about the privacy fears? What of the possibility that the police will use face-recognition systems to conduct illegal searches, or that con men will use them to find victims, or that, more fundamentally, they'll erase our last shreds of anonymity? These are all legitimate concerns, and indications of how much more annoying (and maybe even terrifying) the world will be to navigate in the face-recognition age. But pointing out the potential abuses of this software—and surely there will be a media outcry—won't make it go away. Technology marches forward and ordinary people—people who will be stalked, thrown in jail, or otherwise harassed on the basis of a facial identified—will be collateral damage.
Soon, though, we'll all learn to live with it. Etiquette and even regulations will develop around when it's OK to point your camera at someone and get her name. It's too late to turn back now: If your face and your name are online today, you've already made yourself searchable. Face-recognition systems simply turn the search on its head—instead of looking for your picture by typing in your name, I can now search for your name by snapping your picture. Don't want yourself searchable, period? You can always get off the Internet, or always leave the house wearing a funny hat and a fake nose.
Correction, July 13, 2011: This article originally misspelled David Petrou's last name, repeating a misspelling from the cited article on Tech Talk. ( Return to the corrected sentence.)
Farhad Manjoo is Slate's technology columnist and the author of True Enough: Learning To Live in a Post-Fact Society. You can email him at firstname.lastname@example.org and follow him on Twitter.
Illustration by Alex Eben Meyer.