Ayad Akhtar: AI Ad Tech Is Warping Our Minds

David Beckam
illustration of a page of book text covered with pop-up windows saying "subscribe," "learn more," "buy now" and pixellated arrows
Paul Spella / The Atlantic

This essay has been adapted from a lecture delivered at the Newark Public Library in honor of Philip Roth.

Something unnatural is afoot. Our affinities are increasingly no longer our own, but rather are selected for us for the purpose of automated economic gain. The automation of our cognition and the predictive power of technology to monetize our behavior, indeed our very thinking, is transforming not only our societies and discourse with one another, but also our very neurochemistry. It is a late chapter of a larger story, about the deepening incursion of mercantile thinking into the groundwater of our philosophical ideals. This technology is no longer just shaping the world around us, but actively remaking us from within.

Explore the December 2021 Issue

Check out more from this issue and find your next story to read.

View More

That we are subject to the dominion of endless digital surveillance is not news. And yet, the sheer scale of the domination continues to defy our imaginative embrace. Virtually everything we do, everything we are, is transmuted now into digital information. Our movements in space, our breathing at night, our expenditures and viewing habits, our internet searches, our conversations in the kitchen and in the bedroom—all of it observed by no one in particular, all of it reduced to data parsed for the patterns that will predict our purchases.

But the model isn’t simply predictive. It influences us. Daniel Kahneman’s seminal work in behavioral psychology has demonstrated the effectiveness of unconscious priming. Whether or not you are aware that you’ve seen a word, that word affects your decision making. This is the reason the technology works so well. The regime of screens that now comprises much of the surface area of our daily cognition operates as a delivery system for unconscious priming. Otherwise known as advertising technology, this is the system behind the website banners, the promotions tab in your Gmail, the Instagram Story you swipe through, the brand names glanced at in email headings, the words and images insinuated between posts in feeds of various sorts. The ads we don’t particularly pay attention to shape us more than we know, part of the array of the platforms’ sensory stimuli, all working in concert to adhere us more completely.

Adhesiveness. That’s what the technology aspires to achieve, the metric by which it self-regulates and optimizes. The longer we stick around—on YouTube or Facebook, on Amazon, on the New York Times app—the deeper we scroll, the greater the yield of information, the more effective the influence. We are only starting to understand just how intentional all of this is, just how engineered for maximum engagement the platforms are. In fact, the platforms have been built, and are still being optimized, to keep us glued, to keep us engaged.

Merchants of attention have learned that nothing adheres us to their traps like emotion, and that some emotions are stickier than others. The new and alluring, the surpassingly cute. The frenzied thrill at the prospect of conflict or violence. The misfortune of others. Perhaps most emblematically, the expression of our anger, rightful or hateful. All of this lights up a part of our brain that will not release us from its tyranny. Our fingertips seek it. To say that we are addicts does not capture the magnitude of what is happening.

The system is built to keep us riveted, to keep that neurochemical leak of dopamine steadily coursing, and it operates with a premium on efficiency, which is to say, the platforms optimize for performance based on empirical feedback. An early architect of the ad-tech model writes that the largest monolingual dictionary in the world, the Dictionary of the Dutch Language, has more than 350,000 entries, and yet is

dwarfed by the size of the keyword lists maintained by … search engine marketers. Like a stock portfolio manager, who keeps a set of assets with a theoretical and current price, the paid search manager maintains encyclopedic word lists along with dollar-sign values, and constantly adjusts bids to reflect realized performance.

“divorce lawyer in reno” / cost per click $1.45 / revenue per click $0.90

“nevada cheap divorce” / cost per click $0.75 / revenue per click $1.10

“nevada divorce lawyer” / cost per click $5.55 / revenue per click $2.75

A decade ago, attorneys seeking damages and making fortunes on contingency fees bid up the value of the word mesothelioma as high as $90 per click, making it the most expensive word in the English language. It would be hard to print money faster than these ad-tech auction markets can rack up profits.

This technology is self-regulating by nature—it evolves, like a virus needing a healthy sampling of the population in order to spawn variations. For the tech to be able to tailor and deliver advertising in its various forms, it needs eyeballs. The more of them, and the longer they stay, the more adhesive the platform becomes and the more revenue it can generate.

John Stankey, the current CEO of AT&T, was unusually clear about this prime directive in 2018, as he addressed his new employees at the just-acquired HBO.

“We need hours a day,” Stankey said, referring to the time viewers spend watching HBO programs. “It’s not hours a week, and it’s not hours a month. We need hours a day. You are competing with devices that sit in people’s hands that capture their attention every 15 minutes.” Continuing the theme, Stankey added: “I want more hours of engagement. Why are more hours of engagement important? Because you get more data and information about a customer that then allows you to do things like monetize through alternate models of advertising as well as subscriptions.”

The platforms that churn through content with the greatest velocity shape the emotional responses of consumers almost in real time. Watch a video on YouTube, or like a post on Facebook or Twitter, and you will be offered another, and another, and another. Behind the suggested offerings is a logic of emotional response. The technology is seeking your trigger, whatever draws you deeper and keeps you clicking. Nothing quite does it like outrage. Moral outrage. Those we know are right to hate; those we love because we are united together against those we know are right to hate. This is the logic behind the viral campaigns leading to the slaughter of Rohingya in Myanmar. And the logic of the increasingly truculent divide between right and left in America today. Driven by engagement and the profit that it generates, each side drifts further and further from the other, the space between us growing only more charged, only richer with opportunity for monetization. The cultural clash in America today has more electrical engineering behind it than we realize.

For more than a generation, science-fiction writers and aficionados have speculated about the possibility and imminence of the singularity—that is, the moment when AI will finally eclipse human intelligence. To many, it’s meant the robot capable of thinking, and with an intellect surpassing our own. Let me suggest that digital problem-solving has already surpassed human capacity. Indeed, our advanced societies are now being ordered by a digital matrix of data collection, pattern recognition, and decision making that we cannot even begin to fathom—and that is happening every single successive millisecond. The synergy of data technology, computer-processing speeds and capacity, and an almost frictionless interconnectivity—all of this enables exchange; delivery of services; production of goods; growth of capital; and, most centrally, the endless catalog of our every interface, however glancing, however indirect, with this system’s sprawling and ubiquitous apparatus. The singularity is here—we could call it the era of automation—and its inescapable imprint on our inner lives is already apparent.

In pursuit of what John Stankey called more hours a day, the technology metes out its steady stream of tiny pleasures as the reward for your sustained attention. Touch the screen—respond to the offered stimuli like a rat in an experiment—and receive what some are now calling a dopamine rush.

What follows from this engagement with the devices is an education, in which the system absorbs our responses and begins to shape them. The fetishizing modality of the human unconscious, until now ever-elusive, is given ordinal form as the technology channels the nebulous pull of our proverbial id with Cartesian clarity, our deepest currents of desire rerouted toward the system’s mercantile ends. This careful, unceasing, inhumanly methodical curation of our pleasure principle becomes a larger force in our psyches, as the devices drip-feed us incremental stimulation, which in turn becomes coextensive with the native ground of our very cognition.

We may not notice that there is less and less time passing between touches of the phone. Every 15 minutes? That was so 2018. We’re in 2021, and the urge to reach out for the screen now feels like a rightful impatience with boredom of any sort. But it isn’t that. It’s withdrawal. And from this endlessly recurring neurochemical deficit is born a sense of circumstance and a syllogism that goes like this: Something is wrong if nothing is happening. Something is always happening on this screen. Nothing’s wrong when I’m on this screen. The habit of succumbing to the syllogism—daily, hourly, minutely—charts a course into an undiscovered country of distrust. Distrust of interior discomfort, whatever its texture. Anxiety and uncertainty on the one hand; boredom on the other.

Embedded in this scheme of endless distraction is a deeper logic. The system has come to understand the fundamental value of always reaffirming our points of view back to us, delivering to us a world in our image, confirmation bias as the default setting. This is the real meaning of contemporary virtuality. In the virtual space, the technology combats and corrects our frustrations with reality itself—which defies expectation and understanding, by definition.

I seek. I find what I know. I enjoy this recognition of myself. I am trained over time to trust in a path to understanding that leads through the familiar, that leads through me. “I” am the arbiter of what is real. What is more real than me?

In its basest form—and make no mistake, the baser the form, the stickier the engagement—what we’re describing here is a profound technological support for primary narcissism. We don’t need to know our Ovid in order to understand the perils of all this self-gazing, and yet, we may nevertheless fail to appreciate just how pervasive the social attitudes engendered by this orientation have become.

Self-obsession as a route to self-realization is, of course, not a new idea. American advertising has been foisting this fiction on us for quite some time, exalting attention paid to even our most fugitive and trivial desires, encouraging us to think of their fulfillment as the ultimate purpose of our national politics. But now the scale of the messaging is unprecedented. Technology floods the zone; the waters never recede. In the process, the landscape and its use are entirely remade. The affirming predicate of exhibitionist displays of self-esteem are conflated with instances of political defiance. Self-valorizing anthems abound. “Me” and “my” have been elevated to epistemological categories. And the now widespread misreading of the self’s fragility as resulting not from the contingent situation of selfhood itself, but from society’s failure to protect and recognize “me.”

Accustomed to gossamer-thin ad-and-subscription-supported satisfactions, absorbed and convinced by the moralizing rhetoric that passes off our dependence on technology as righteous activism, we internalize another pernicious untruth, deeply damaging to our social fabric—namely, that the path to redemption and change will be paved by our personal pleasure, pleasure we come to feel we shouldn’t have to suffer even a moment’s discomfort in order to enjoy. To use a beloved locution borrowed from the lexicon of contemporary self-esteem culture, we deserve this pleasure because we deserve better; we deserve to feel good.

All of this points to a new social ontology, an evolving set of behaviors guided by the shift in incentives that the technology has created. It’s the advertising model of thought; the entertainment model of consciousness. Self-promotion, self-commodification, self-marketing—all are now taken for forms of legitimate commentary and critique; ceaseless affirmation of our biases emboldens the strident certainty of our moral positions. This is the complexion of public exchange in the newly shaped public sphere, where ideas are little more than bait for the hours a day of human attention at stake, yet another demonstration of just how much the technology is reshaping our relations with one another. In fact, we are little more than grist for a monetizing mill that mixes, like cattle feed ground from cattle bones, our own deepest intimacies with the system’s digital slop, feeding it back to us wholesale. In the process, we are being remade by what we consume. In the writer Chimamanda Ngozi Adichie’s words: “I notice what I find increasingly troubling: a cold-blooded grasping, a hunger to take and take and take, but never give; …an ease with dishonesty and pretension and selfishness that is couched in the language of self-care; an expectation always to be helped and rewarded no matter whether deserving or not; … an astonishing level of self-absorption; an unrealistic expectation of puritanism from others; an over-inflated sense of ability, or of talent where there is any at all; an inability to apologize, truly and fully, without justifications; a passionate performance of virtue that is well executed in the public space of Twitter but not in the intimate space of friendship.” Stirring words. I doubt anyone will fail to see some truth in what she’s saying. But perhaps even more disturbing than the pain behind this passionate indictment is the predicament of those who occasioned it. The rest of us. Because who, if they’re truly honest, would dare think they’d somehow escaped?

In the end, a writer survives only if there’s wisdom in their work.

A hundred years later, a reader has to recognize the emotional patterns as their own, no matter what the social circumstances of the writer was.
— Vivian Gornick

As a writer, it seems to me that the most baleful development in our collective contemporary life is the preponderance of a practice derived from digital technology that treats knowledge and information as synonymous. For while the way to wisdom leads through knowledge, there is no path to wisdom from information. Especially when that information is being used as a training treat in what has come to feel like a wholesale attempt at permanent reeducation.

Having one’s bias confirmed endlessly by a curated cascade of information reflecting back to you your preferences and opinions, second after second, understandably breeds an illusion of certainty. But certainty is nothing like wisdom; it might in fact be something closer to wisdom’s opposite. Wisdom: a kind of knowing ever-riven with contradiction, a knowing intimate with the inevitability of uncertainty. In American Pastoral, Philip Roth writes:

Getting people right is not what living is all about anyway. It’s getting them wrong that is living, getting them wrong and wrong and wrong and then, on careful reconsideration, getting them wrong again. That’s how we know we’re alive.

A commitment to recognizing that one is wrong, to the necessity of failing to come to a conclusion and needing to make sense again, and again, and yet again—this is the path to what Gornick identifies as that which endures in literature. For Saul Bellow, novels were tests of his best ideas, in which he hoped, indeed expected, that those ideas would ultimately fail. Certainty is anathema to art, to wisdom, and ultimately, as Roth writes through the voice of Nathan Zuckerman, to living itself.

The regime of screens afflicting our cognition has elevated the centrality of certainty in our public spheres, which are increasingly indistinguishable from our private ones. And even more alarming for the writer, who cannot trust certainty as a guide to her work, it is pilfering our elective affinities, transforming these into ones selected for us. By using the term elective affinity, I am invoking an idea that runs from alchemy through Goethe and into the social sciences of the 19th century. The notion being of chemical elements, or people, or cultural forms that evince analogy and kinship, and which, therefore, enter into mutually arising relationship. For a writer, affinity is the guide, the lamp that lights the way.

It was ever so for Roth, a writer of fearless and passionate affinities, if there ever was one. Affinities he felt toward the great American writers, even at a time when the prevailing social thinking imagined him as Jewish first and American second, if at all; affinities that led to his embrace of even a virulent anti-Semite like Céline. “Céline is my Proust!” Roth once said. “Even if his anti-Semitism made him an abject, intolerable person. To read him, I … suspend my Jewish conscience … Céline is a great liberator.” The path of affinity often leads to contradiction, like that of an American Jewish novelist emulating an avowed anti-Semite. Contradiction, which, if F. Scott Fitzgerald is right about the test of a first-rate mind, is but another shape that wisdom takes.

In Benjamin Taylor’s Here We Are, a touching account of his friendship with Roth, Taylor writes of Roth reading aloud a passage in Joseph Conrad’s Lord Jim:

A man that is born falls into a dream like a man who falls into the sea. If he tries to climb out into the air as inexperienced people endeavor to do, he drowns. The way is to the destructive element submit yourself, and with the exertions of your hands and feet in the water make the deep, deep sea keep you up … In the destructive element immerse.

And Roth looks up and adds: “It’s what I’ve said to myself in art and, woe is me, in life too. Submit to the deeps. Let them buoy you up.” The downward movement that lifts. Or the ascent that sinks “slowly as a kite”—as Elizabeth Hardwick writes in Sleepless Nights. The paths to the wisdom of contradiction are legion, which is why any artist with a nose for a possible route, alive to her own affinities, will not ultimately be bucked by the concerns of the many.

Whether the concerns of the many are louder today than before is hard to know. But they may be more inescapable. One of the characteristics of the automating technology is how effective it is in herding opinion in ways not meaningfully different from policing it. The tech has created gathering places for our various camps of confirmed bias. These agglomerations of outrage are not just left-leaning or right-leaning, groupings superintended by slogans of belonging and creedal statements honed, like trademarks—or shibboleths—to the very locution. The result is a widespread and punitive stridency.

The writer today, wherever she is, must not be cowed by fear, however real, of opprobrium, retaliation, and group exclusion. She must know that the path to the transmutation of knowledge that produces the wisdom of literature can, in the end, lead only from her own sense of things. Information masquerading as knowledge will not lead her there, nor will the metaphysics of group belonging redeem her in the form of monetized moral certainty. Attention paid to her own affinities, however heterodox, may be all she has to go on, and the fear of doing so is what she must defy. Which is why any defense of the path to literature, to the writing of it—and by extension, to the reading and teaching of it—can only be as strong as those willing to heed it. Fundamentally, this is not a matter of judgment, not for a court of public opinion or of any other sort. It is a matter of the heart, a matter of that wisdom we call love.


This article appears in the December 2021 print edition.

https://www.theatlantic.com/magazine/archive/2021/12/ai-ad-technology-singularity/620521/

Next Post

Toyota dealers: Improving customer experience a top priority

How deep are you selling into your pipeline? Right now, of my F-status vehicles, or freight status vehicles — the vehicles coming at me — 95 percent of them are sold, and they’ll all be sold probably within a day as they keep coming in. And we’re selling well into […]

Subscribe US Now