Earlier this month, Facebook launched an aggressively adorable stand-alone photo app called Moments. Though almost offensively cheery, the app is actually pretty cool, allowing multiple users attending the same events to privately cluster and share photos into a secure communal pool — great for weddings. Even more so for bachelorette parties.
Cuteness and cleverness aside, the real reason Moments may soon experience something of a moment is as a potential outlet for Facebook’s highly refined facial recognition capabilities, which you’ve likely experienced if you’ve ever tagged a photo — or had one suggested for you.
Facebook is still hard at work developing DeepFace, a powerful mega-algorithm (which Facebook presented last year at the Conference on Computer Vision and Pattern Recognition in Boston) that can factor in a wide range of biometric data (like the shape of your body and hair, or the clothing you wear) to reach something like 97 percent accuracy, or “approaching human level performance.” The advances of DeepFace will address a number of fundamental shortcomings of facial recognition — the kind of limits that made it unsuccessful in identifying the Boston Marathon bombers.
Meanwhile, the level of facial recognition employed by Facebook is already advanced enough that the Moments app has been blocked by European regulators, whose resistance to cuteness and cleverness extends to potentially invasive technologies that don’t allow users to opt in.
Facial recognition is doing its best to put on a friendly face — it’s recently helped reunite a kidnapped twin with his family, and one new app will employ it to locate lost dogs — but it can probably detect the skeptical sneers coming its way.
There’s a built-in creep factor to facial recognition, a strong whiff of Orwellian/K.-Dickian nightmare. The thought of a massive archive of our most personal possessions drained of their humanity naturally gives many of us the willies (as well as flashbacks to the “Game of Thrones” finale).
The FBI maintains a massive database of facial data, with aims of over 50 million faces on file, and 85 percent accuracy (meh, they’re no Facebook). These systems can also identify you through “soft biometrics” like your height, size, gait, even the details of your ears.
Companies like the California-based FaceFirst are engineering systems for law enforcement, security (from airports to casinos), and retail use. One platform, Churchix, is designed specifically for tracking church members’ attendance — presumably greatly freeing up God’s schedule.
And as these technologies surge forward and spread outward, there are mounting legal questions about how they’re used, and more importantly, how we consent to that use. One lawsuit underway in Illinois against Facebook (and a similar case in that state against Shutterfly) accuses the service of running afoul of state “biometric privacy” statutes (laws that remain unique to Illinois and Texas).
And federal laws don’t yet address the concerns of facial recognition. This past week, nine groups of privacy advocates walked out of talks with the National Telecommunications & Information Administration, which sought to “develop a voluntary, enforceable code of conduct that specifies how the Consumer Privacy Bill of Rights applies to facial recognition technology in the commercial context.” The group issued a statement on the impasse, complaining that “in recent NTIA meetings . . . industry stakeholders were unable to agree on any concrete scenario where companies should employ facial recognition only with a consumer’s permission,” and that “people deserve more protection than they are likely to get in this forum.”
Amid all of this uncertainty and anxiety surrounding the hazards and implications of facial recognition, I’m left with one question: Why do I still think it’s so cool?
If there’s one thing that’s not surprising about the development of advanced facial recognition technology, it’s that we developed advanced facial recognition technology. It’s the most obvious advance of them all, implicit in every past attempt at convenience (the ATM that recognizes you as you stroll up is truly an automated teller) and built into our understanding of how humans will eventually interface with AI.
Maybe I’m just selfish and tech-a-dent. I want the store with my weird jeans size to tell me it has my weird jeans size in stock. I want the robo-concierge to effortlessly consult the virtual nametag of my mug. I’d like to get on a plane faster; for Panera to know about my nut allergy; for my face to be as powerful as my phone.
None of this is to discount the importance of privacy, but part of our resistance to recognition is caught up in our valuation of the face as something greater than mere data. The sensitivity we feel toward pictures of our face (from our dashed-off drivers’ licenses to our highly regulated selfies) is not something we extend to, say, our equally-as-individual fingerprints — their tiny contours may be inscrutable and alien to us, but not to the scanner at the front counter of the gym.
For now, when it comes to facial recognition, the matter of comfort is wrapped up in the question of consent; and it’s unclear if that tension will be resolved one affirmative click at a time, or en masse, as part of a cultural shift in how closely we identify with the Internet. Privacy advocates are already coming up with creative ways to dodge the gaze of the mainframe — from donning fashionable scanner-thwarting “camouflage” accessories designed by CV Dazzle, or by varying your appearance enough to blur your “faceprint” (good news, Rachel Dolezal).
But while it satisfies every paranoid impulse within me to resist or recoil from a technology as loaded with potential as it is haunted by dystopian dread, there’s just as much of me inclined to face it: Smile, say “cheese,” and see where all of this can possibly lead. How, without facial recognition, will Rosie my housekeeping robot know who I am when I walk in the door? And what will move her to ask if there’s something troubling me?