This essay was originally published on September 22 2018.
Public media discourse has been fairly agonizing. All sides seem beset by mobs of malicious morons: waves of fake news pounding against their believers, crafty hackers seeding chains of disinformation, the trollish alt-right cooptation of identity politics, and even a populous turnout at the US midterm elections that, no matter your political leanings, likely left you with mixed feelings. None of these phenomena can be written off as pure idiocy, however. After all, successfully spinning the news, hacking, trolling, and campaigning are all sophisticated and strategic crafts in their own right. It takes a lot of cleverness — even cunning — to render media environments so excruciating. How can so much of public talk feel so dumb and so smart at the same time?
Might intelligent tech, once examined, offer a surprise insight? Consider the case study of “smart media” — understood here as smart devices such as the wifi-enabled interactive phones, watches, speakers, doorbells, homes, cars, and even cities, as well as the popular enthusiasms for deep machine learning, neural nets, and artificial intelligence. What exactly do we mean when we call media technologies “smart”? What should “smart” mean?
Intelligence is both the holy grail and the curse of the modern media environment. Everyone wants it in infinite supply, but no one knows exactly where to find it. Easy definitions evade specialists: one study in 2007 collected over seventy definitions of “intelligence,” finding only broad-brush overlaps between them. For example, psychologists measure “comprehending our surroundings, ‘catching on,’ ‘making sense’ of things, or ‘figuring out’ what to do” with the seemingly modest lower-case metric “g” — the intelligence factor lurking behind IQ tests — as a complex composite of many other specific skill tests. Artificial intelligence researchers, too, often define intelligence as an aggregate gauge of how well an agent achieves its goals relative to others.
But there is something unsatisfying about this understanding. For one, if intelligence just means outperforming one’s peers, a know-nothing real estate tycoon who becomes president may be surprisingly justified in declaring himself a “very stable genius.” It is a despairing thought that intelligence might be little more than a letter grade awarded to differentiate peers on a thoroughly uneven playing field. Instead it might be wise to take a page from the philosopher Thomas Hobbes, who (except for mentioning government intelligence gathering in the Leviathan) avoided the term altogether: he, for example, rendered the logical absurdity in Latin intellectus intelligit as “the understanding understandeth.” If intelligence is what an intelligence test tells us it is, then perhaps classical thinkers have, with good reason, dwelled on education, experience, judgement, reason, and wisdom instead.
Public life, alas, has not been spared such luxury: “smart” things continue to proliferate. Estimates of the number of networked devices likely to wire the world by 2025 range in the billions, tipping toward a trillion. The consulting firm McKinsey estimates an annual growth rate of 31% in the smart home device industry in the last half of this decade. Smartness may be expanding, but wisdom isn’t. Smart home tech has proven (in a list curiously reminiscent of the right’s criticisms of liberal college students) wildly oversensitive, vulnerable to whims, and even cruel: recent reporting in the New York Times has uncovered smart domestic abuse, where (almost always) men remotely control the women in their lives, using smart device controls to lock doors, turn off and on lights, and change temperatures to torment their victims from a distance.
As different types of machine intelligence advance, we see reflected still other limits of our social intelligence: despite significant advances in algorithms that perform highly-constrained tasks such as besting human experts in games like poker, chess, and Go, image recognition, and other forms of machine learning, our most successful algorithms are also now learning and amplifying our human biases — classist, sexist , racist, etc. Those who imagine that these biases reflect a solvable flaw in our technology, rather than in ourselves, are likely to make the mistake of seeing strong artificial intelligence as a science rather than the speculative philosophy that it is. There is perhaps no folly more modern than to pine after artificial intelligence while neglecting the real pains caused by the artifices of human intelligence.
A signal lesson: “smart” is not the opposite of stupid. In tech a “smart” device refers to a different infrastructural category than in humans: smart specifies a device that is wifi-enabled, not steeped in the chemical bath of neurons and synapses. Now consider that the very last thing most people would want to be “smart” — in the tech sense — is one’s own brain, the conventional seat of human intelligence. (The exception proves the rule: at least one colorblind cyborg artist has wired an antenna into his head to let him hear colors.) What does it reveal that tech “smartness” and the center of human intelligence seem so incompatible?
Smart and stupid are not opposites in another broader, deeper, and more abiding sense: “smartness” may have less to do with intelligence than it does with its verbal definition: that which smarts stings us with cutting intensity and bodily pain. In English, one might rub a bruise while blurting out, “ow, that smarts” — a term that shares a telling root with “Schmerz,” the German word for pain.
May we understand smart media as painful media — or media that prompt us to account for our bodies, their limitations, and how technology extends our embodied experience into new spheres of experience that sting with cutting pain. A smartwatch loops the user — as well as anyone else with access — into tracking their movements; while a smart oven promises to extend to great distances the reach of culinary creativity — including the risk, among others, of burning down the house. Such an approach can help refresh and reclaim the critical potential that “smart” does in contemporary media and technology talk. Common to all these smart devices and their discontents — the everyday inconveniences and failure rates of fragmented tech, the remote-control abusive cruelties they enable, the algorithmic biases against the disadvantaged and disenfranchised, the cunning trolls and mainstream identity politics hackers—is the fact that their exploits depend on leveraging the painful, personal, and physical realities of embodied humans and our tender inequalities. Examples: “lulz”—the success threshold for a troll — is best obtained through the salty tears of the target of their attacks; as critical media scholars have shown, algorithmic bias invariably amplifies social biases against real-life embodied differences — ethnicity, sex, education, disability, etc.; and remote control over victims of smart device abuse takes place not at the remove of “online,” but at home — in the most intimate of spaces.
It seems smart devices are adopting a new kind of agency: things appear less and less on the internet, and more and more is the internet in things, even in us. Or rather, the internet of things is increasingly the internet of our appendages, and the cephalopod-like reflections of “smartest” ourselves in the network mirror are sure to make us uneasy.
And so they should. This basic revelation — smart media is a pain — deserves another hard look: as I’ve argued elsewhere, the word “digit” referred to fingers and hands for hundreds of years before it described computing. By analogy, so too do smart media reground us in our bodies and their extensions. Our media environment feels so painful because it does smart. It smarts: it senses, surveils, and often hurts us where we live. Even as investors and advertisers continue to pedal ever newer and tech “smarter” ways to automate away the labor of human behavior, we must acknowledge that smart media — true to its name — will continue to reflect and amplify the only and often-excruciating “broken-world” reality we have: living out together our fragile lives in mortal bodies on the finite blue wet ball we call the planet earth.
In a media age overripe with virtual disembodiment and escapist promises, attending to what really smarts just might be the smart thing after all.
Benjamin Peters is the author of How Not to Network a Nation: the Uneasy History of the Soviet Internet (MIT Press 2016) and editor of Digital Keywords: A Vocabulary of Information Society & Culture (Princeton University Press 2016). He tweets @bjpeters.