Faces and names
A few years back I used to post images like the one above every now and then to Facebook. I used to make them at work in InDesign between copyediting news articles. I had no particular agenda; it just seemed to make sense to me, a literal way of turning myself into content. Manipulating images of my face felt like a way of disassociating from it — the only way to make looking at myself feel tolerable. It made my face feel more low-stakes, less significant, less like a trap I was inevitably caught in.
Posting these images to social media intensified that feeling of escape, as though I was getting rid of my face by casting it out onto the feeds. It let me feel as though my face wasn't "who I am" but instead just something I appropriated and reused, like any other meme. I could parody it, or deploy it as obscure commentary, or layer it on itself like some glitched logo.
At the time, I had this idea about selfies being less a matter of trying to represent oneself and more a way of demonstrating your availability to the network — it showed a willingness to construct (or be constructed as) a series of alienated self-representations for the benefit of social media platforms, an eagerness to be circulated and multiplied as an image. The "self" in a selfie, I thought, is not what is recorded in the image so much as how it travels and what other people do with it in response. No matter how calculated and perfected a selfie might be, one cedes control by posting it — but one also leaves that idea of the self behind.
I still think of selfies that way, as a way of getting rid of yourself or multiplying versions of yourself, letting a certain rendition of your identity travel without you. Every selfie implies a series of selfies; you are stuck with one face but can generate a countless number of selves, each canceling or complicating the others. The act of posting a selfie lets you move on from that self while it moves around online.
Even if they aren't posted, selfies can serve this sort of self-alienation function: In The Selfie Generation Alicia Eler notes how "selfies that are shot but never posted are simply part of an individual's momentary self-reflection" — the fact that people often have "lots of selfies in their phone photo album" serves as evidence of how often people elect to think about themselves as objects.
It would seem like facial recognition technology — which takes translates your face into a data set makes it commensurate with all other faces, and allows machines to play with it — is an extension of this impulse to demystify the face, make it something to fool around with. But in letting a facial recognition system boil us down to a pattern — that creepy vector grid that used to overlay your face when an app had successfully scanned it — it seems as though there is a tendency to surrender to that system, to abdicate the agency over our self-presentation.
In an essay for the New Yorker about the Google app that matches your face to a work of art, Adrian Chen described the uncanny ways in which we pose for photos that will know will be viewed mainly by machine vision: "As my social-media timelines filled with images of my friends with their doppelgängers" — the diptychs the app makes of a user's face and a painting — "I was struck more by their photos than the matches. They were often alone, poorly lit, looking straight into the camera with a blank expression ... Unlike the well-composed selfies and cheerful group shots that people usually share, these images were not primarily intended for human consumption." He concludes by reversing the point often made about training data sets — whether or not Google is collecting images of our faces to train its algorithms, we are certainly downloading facial recognition apps to train ourselves in how to be recognized. To be read by a machine, we almost instinctively try to empty ourselves of subjectivity, present ourselves in a formalized, neutralized way.
The way Apple is selling the face-scanning technology on its iPhone X is similarly muddled: It's supposed to be superfun to unlock a phone with your face, as if this confirmed your secret fantasy that your face was the key to the universe. But in practice, this means positioning your face so that it can be read like a bar code. It requires you to habituate yourself to becoming a bar code, assuming the "blank," flat, unambiguous posture of data. You present yourself to be counted, to be scanned like a package.
Likewise, it's supposed to seem creative and fun to use your face to operate a talking turd emojj, but this comes at the expense of the expressive range of the face in its own right. The face is turned into a control panel, an overlay, no more or less complex than an Instagram filter — it's just a static data set that can be imposed on any kind of content like a rubber stamp.
The latest iPhone seems to see all faces as unique — as verifiably ours and ours alone — but this is because it sees all faces the same way. It imposes the same idea of what a face is on everyone and merely calculates everyone's variance from the standard — that is, it is inherently normative. It treats the face as if it were fundamentally a unit of identity and nothing more, allowing that function to overcode all other possibilities. Whatever you can express facially, the most important thing is that it's verifiably you expressing it.
If the face also conveys something about what one is trying to express or become, or any sort of reaction to what is going on, that is merely incidental. You are just the same number everywhere, whether you are face to face with a lover or just a face in the crowd. Even though you might look different to people in these situations, technology knows better. Only your phone can recognize you.
Jacob SIlverman, in this Baffler essay, stresses the point that facial-recognition technology treats the face as a biometric marker — a digital hash that can be reverse-engineered or hacked. "You can change a hacked password," he notes, "but not a face." It's like nature's blockchain.
Facial recognition systems bind individuals to a ID marker that can be scanned at long distances by surveillance cameras, so one's presence may in theory always be detected, tracked. Moreover, this imposes a sense that you remain the same person, the same self, no matter where you go, for whatever reasons — that there is no reason outside agencies should not blend all this data and draw conclusions about what you really want or what you are really about. The spaces in which you might share a selfie to discard an identity, to shed an old idea of yourself, or proliferate many selves, are closed off. At the level of code, facial recognition militates against the supposed duplicity of going by many names, of having more than one identity. Your face is a static "real-name profile" that is read off of you against your will — it imposes a single identity that is in the control of the recognition system and guarantees you can't escape it.
But no one actually lives by such standards: We become different people in different situations and different places with different people. We are called by different names. We are constantly impersonating others, or ideas, or roles, or other people's ideas of ourselves or a role. We blend into groups, become part of collective subjectivities.
Facial recognition technology denies the existence of such a thing: anything with a face is always consigned to full identity with itself, and full responsibility for itself. Because you can be tracked individually by your face, the presumption is that agency operates the same way: You can now be held accountable as an isolated individual for anything that occurs where you are. There won't be any need for anyone to name names.