“He sees you when you’re sleeping. He knows when you’re awake. He knows if you’ve been bad or good, so be good for goodness sake!” – “Santa Claus Is Coming to Town”
When we worry about surveillance, we do so from the part of us that grunts, not the part that fusses about rights. The fear of being watched is blood-borne — the Original Fear, the prospect of predation. George Orwell understood this. His “1984” is more than a political fable or a swipe at utopian ideals — it is a horror story. And reading the book, you get the sense that the nightmare does not reside in Big Brother, the unblinking eye of the State. The scariest thing about the tale is this: The real predators are your co-workers, your neighbors, your actual big brother.
The snitch culture portrayed in “1984,” of course, predated Orwell (and the Communist despots he loathed), and it outlived him, too. Osama bin Laden met his end at the hands of a tattletale. Two thousand years ago, so did Jesus Christ. But the last few weeks have shown us something new. Suddenly, without fanfare, we have entered the age of the Super Snitch, and its name is Facebook.
We caught a glimpse of this mythical-beast-to-be during the recent UK riots. A strange thing happened to Trev, Kev, and their fellow looters on the way to their local Xbox retailer: To get there, they had to elbow their way through small clusters of concerned citizens, each of them wielding a smartphone, quietly clicking. You could see these so-called digilantes on the news footage, their arms extended at about chin height; they looked a little like vampire hunters, digital crucifixes in hand.
The rioters, seeing them, would probably have been more confused than concerned. It probably hadn’t occurred to them that images of their window-kicking exploits would end up alongside their profile pics — Kev sipping alcopops in his local pub, Trev headlocking his girlfriend on a beach in Marbella. Had they paused to consider Facebook’s new Tag Suggestions feature, which uses state-of-the-art face recognition software to automatically put a name to the faces that appear on the site, they might have stayed home and watched the action on TV.
As the violence intensified, a Google group cropped up, London Riots Facial Recognition, a fellowship of techno-geeks who offered to develop a system to help put marauders behind bars. While this group of do-gooders eventually petered out — digilante-related arrests have been few so far — thousands of incriminating shots are now pinging around the social networking universe, subject to an endless, irreversible process of clarification.
It may not happen today, Kev, or tomorrow, Trev, but it will happen: Your faces will give you away.
“A man finds room in the few square inches of the face for the traits of all his ancestors; for the expression of all his history, and his wants.” — Ralph Waldo Emerson
The face is a kind of logo. It says who we are, and not just in terms of identity. The face can contradict us. It can betray us (Nine-three, it’ll chime after a potentially catastrophic all-in at the poker table; nine-three, unsuited). The face is the most intensely personal feature of the human anatomy. If we thought about it too much, we’d all be walking around with underpants on our heads.
Face is character. It is also, now, date of birth, e-mail address, best-ever movie, and what you had for breakfast. Today, based on nothing more than your so-called faceprint, random strangers can learn all sorts of interesting things about you. Not the traits of your ancestors, perhaps, but pretty much everything else. And, again, it’s not the State that’s doing the watching here, it’s the guy you passed on the street today, seemingly engrossed in something on the screen of his device.
Over the past year or so, the big players in recreational digital technology have been quietly trawling the research community, looking to nudge ahead in what might be called the Face Race. Polar Rose, a software developer snapped up by Apple last year, has in its portfolio Recognizr, an app that allows you to point your phone at a random individual and click. If your target has the app, too, you will see a bunch of icons orbiting his or her head, like little planets. These may be social networking sites, e-business cards, iTunes preferences, anything the user has chosen to include. Touch one of these icons, and voila, instant familiarity.
This summer, Google acquired Carnegie Mellon spin-off PittPatt, a developer of eerily perceptive face recognition software. While the technology is ostensibly intended for use on the Google+ networking site, a way for the company to nose in on Facebook’s photo-sharing business, there are other avenues for it to take. Google, after all, already has Goggles, a remarkably efficient image recognition app that allows you to take a picture with your phone, then search the Web for information about the photo.
Facebook, meanwhile, is said to have employed the know-how of Face.com, which has promised to put a name to every picture on the Internet. Face.com has recently been tinkering with the principles behind Microsoft’s Kinect technology, which allows video game characters to mimic the expressions and movements of players, developing software that, it claims, can gauge the mood of a person based on the twitch of an eyebrow, the curl of a lip.
These technologies will merge and mutate. We are on the verge of something very strange.
3. The Fog
“I never forget a face, but in your case I’ll be glad to make an exception.” – Groucho Marx
The topography is deceptively simple. We are defined by tiny variations in the slope of the nose, the angle of the jaw, the set of the eyes. Taken alone, these elements are like the individual digits of a telephone number — meaningless until put together in a certain way. What face recognition software does is create mathematical algorithms out of this tiny constellation of features, or nodes, then match these with identical patterns that have been stored in a database. Easy.
Yet, in the 40-odd years that the technology has been in use, face recognition has consistently failed to connect the dots. After 9/11, Boston’s Logan Airport installed a system whose accuracy rate was a little over 60 percent — placing it in the same bracket as a doddery, though not quite senile, socialite. “Harry!” it would say. “No, wait . . . Henry!” And its eyesight was as prone to failure as its powers of recall. “Hosni?”
But the machines are getting better. Last month, a group of German researchers announced that they’d employed face recognition software to identify individual chimpanzees at the Leipzig Zoo, with an 83 percent success rate.
With reliability on the up, so is usage. This fall, police departments across the United States will be equipped with snappy hand-held devices that allow officers to scan the face of a suspect and produce an instant criminal profile — outstanding warrants, prior record, aliases, etc. In the United Kingdom, sophisticated technology developed for the 2012 Olympics has been used to sift through the mountains of CCTV shots the riots generated. London’s Heathrow Airport is deploying recognition software which claims a maximum “False Acceptance Rate” of 0.0125 percent, which is about as close to infallible as the technology has gotten.
Oddly, perhaps, privacy advocates have been relatively mute with regard to the security and law enforcement applications of the technology, opting instead to work themselves into a lather over the say-cheese sphere of social networking — specifically, Facebook’s new auto-tag feature, which has, in recent weeks, become the focus of an official investigation in Europe, an FTC complaint in the United States, and the brow-furrowed blathering of tech bloggers around the world.
There’s a sense, though, that the people targeting Facebook are swatting at snowflakes, oblivious to the drifts gathering around them.
“And it must follow, as the night the day, thou canst not then be false to any man.” – William Shakespeare, “Hamlet”
We’ve all seen it, the grainy footage of mankind’s early efforts at flight — wings snapping upwards, noses furrowing the earth. Skip forward a generation or two and we have images of Amelia Earhart and Howard Hughes, sleek and silvery aircraft at their backs — this is about where we are with face recognition, the breakthrough point. Until recently, however, the technology has been lacking a key ingredient, something that’s hard to put your finger on — life, maybe, or experience.
When humans see a familiar face, we immediately start making associations, we try to place it. We met this guy the night we choked on an olive; he looked a little like Alan Rickman; he told a joke nobody laughed at; he kept making kissy lips at our newly divorced friend. The process is not a matter of analysis; it’s a tumble of synthesis, intuition, and improvisation, and matches are often made in ways that even the most advanced software has never been able to emulate. Until now.
Ironically, given the amount of time and money that’s been poured into the enterprise, the big leap forward in artificial face recognition has not been brought about by design. The difference now is not that we have access to greater computational power or more sophisticated algorithms. The difference now is you — or, more specifically, you and your big mouth.
We routinely, unthinkingly seed the blabosphere with details of our plans for the weekend, our arguments with lovers, our awkward illnesses, our super-hot co-workers, our aversion to spinach. With all this information at its disposal, this virtual infinity of personal trivia — along with, of course, the countless attendant photographs — face recognition technology is able to delve around for associations in much the same way people do. Social networking sites have, in this regard, begun to emulate the realm of human experience.
Which brings us to the unfortunate thugs mentioned above, Kev and Trev — facing imprisonment because two random, unrelated images were automatically linked, as if some kind of digital synapse had fired. It’s no accident that, in the aftermath of the UK riots, the police could be seen rooting around on sites like Flickr. They have the technology, they have the digital mugshots, but until recently they didn’t have access to our collective memory.
George Orwell would have been morbidly fascinated by all this. The hackers and stalkers will be interested, too. The pornographers, debt collectors, and direct marketers are standing by. The jilted lovers will have their day. And the guy you passed in the street today — well, he’s back in his bedroom, hunched over his monitor, adding you as a friend.
Chris Wright is a writer living in Spain.