In fact, Varian’s notion of personalization is the precise opposite of the relationships with trusted professionals to which he refers. Doctors, accountants, and attorneys are held to account by mutual dependencies and reciprocities dictated by the extensive institutionalization of professional education, codes of conduct, and procedures for evaluation and review. Violation of these rules risks punishment in the form of professional sanction and public law. Google and its brethren in surveillance capitalism bear no such risks.
Varian’s remarks constitute one of those rare occasions in which the fog of technology rhetoric parts just enough to discern the utilities of social and economic inequality to surveillance capitalism’s larger mission. Varian reasons that inequality offers an opportunity to raise the ante on Google’s quid pro quo for effective life. He counsels that the way to predict the future is to observe what rich people have because that’s also what the middle class and the poor will want. “What do rich people have now?” he asks rhetorically. “Personal assistants.”
That the luxuries of one generation or class become the necessities of the next has been fundamental to the evolution of capitalism during the last five hundred years. Historians describe the “consumer boom” that ignited the first industrial revolution in late-eighteenth-century Britain, when, thanks to visionaries like Josiah Wedgewood and the innovations of the early modern factory, families new to the middle class began to buy the china, furniture, and textiles that only the rich had previously enjoyed. This new “propensity to consume” is considered “unprecedented in the depth to which it penetrated the lower reaches of society.…”3 In 1767 the political economist Nathaniel Forster worried that “fashionable luxury” was spreading “like a contagion,” and he complained of the “perpetual restless ambition in each of the inferior ranks to raise themselves to the level of those immediately above them.”4 Adam Smith wrote insightfully on this social process, noting that upper-class luxuries can in time be recast as “necessaries.” This occurs as “the established rules of decency” change to reflect new customs introduced by elites, triggering lower-cost production methods that transform what was once unattainable into newly affordable goods and services.5 Ford’s Model T is the outstanding twentieth-century example of this progression.
Varian casts personalization as a twenty-first-century equivalent of these historical dynamics, the new “necessaries” for the harried masses bent under the weight of stagnant wages, dual-career obligations, indifferent corporations, and austerity’s hollowed-out public institutions. Varian’s bet is that the digital assistant will be so vital a resource in the struggle for effective life that ordinary people will accede to its substantial forfeitures. “There is no putting the genie back in the bottle,” Varian the inevitabilist insists. “Everyone will expect to be tracked and monitored, since the advantages, in terms of convenience, safety, and services, will be so great… continuous monitoring will be the norm.”6 Everyone, that is, except those wealthy or stubborn enough to achieve effective life without Google’s assistance and thus escape the worst excesses of rendition. As decision rights and self-determination become privileges of the wealthy, what will Varian offer to answer those who clamor for the same?
Historically, breakthroughs in lower-cost goods and services unleashed expansions of production and employment, higher wages, and an improved standard of living for many. Varian has no such reciprocities in mind. Instead, he pokes his finger into the open wound of second-modernity insecurities and bends our pain to the objectives of the surveillance project. With Varian, the hunger for new necessities is construed as an opportunity to dispossess, even as it conveniently provides the justification for that dispossession right down to the depths.
Google Now was a first step, although later it would look more like a stalking horse and habituation exercise paving the way for what was to come. Christened “predictive search,” it combined every system that Google had ever built, including the corporation’s achievements in voice search and neural networking, its knowledge of the world represented in its one-billion-entity “knowledge graph,” and its unparalleled machine intelligence capabilities. All of this firepower was amassed in order to learn from your content, context, and behavior not only through search, e-mail, and calendar activity but also from the data in your phone, including movement, location, activities, voice, and apps. This time, the aim was not only to sell ads but rather “to guess the information you’ll need at any given moment” as you move through the real world.7
As a promotional video crows, “Google Now is always one step ahead so you can feel more confident as you navigate your day… with the predictive power of Now, you get just what you need to know, right when you need it.” One writer described the new service as “having the search engine come to you.”8 The app’s information cards swim into view on your phone’s home screen in anticipation of your needs: notification of a changed flight time, impending weather and traffic, nearby restaurants and shops, that museum you’ve been wanting to visit. One Google executive reasoned that Google already knows all of this about you, so it might as well turn it into a service that can provide the company with access to even more information: “Google’s going to know when my flight is, whether my package has gotten here yet, and where my wife is and how long it’s going to take her to get home this afternoon.… Of course Google knows that stuff.”9 Google Now’s predictive capabilities follow the pattern we have seen throughout: they derive from machine processes trained on unceasing flows of virtual and real-world behavior. Why did Google devote so much machine power and valuable surplus in order to thoughtfully assist you through your day? The reason is that Google Now signaled a new breed of prediction products.
Google’s breakthrough crawler enabled the lightning-fast indexing of the world wide web, the apparatus of ubiquity then enabled new operations to crawl reality, and now in this third phase, distinct supply operations are required to crawl our lives. In Google Now one sees an initial foray into this new space, in which the web crawler’s ability to find information combines with new life-crawling operations intended to render, anticipate, and, as we shall see, ultimately modify your behavior. Online and offline behavioral surplus—your e-mail content, where you went this afternoon, what you said, what you did, how you felt—are combined into prediction products that can serve an emerging marketplace in which every aspect of your daily reality is up for bid.
Facebook’s “M,” launched in 2015 as part of its Messenger application, is another example of this new phase. It was introduced as a “personal digital assistant… that completes tasks and finds information on your behalf… powered by artificial intelligence that’s trained and supervised by people.”10 Facebook’s vice president in charge of messaging products described the company’s goals for M by saying, “We start capturing all of your intent from the things you want to do. Intent often leads to buying something, or to a transaction and that’s an opportunity for us to [make money] over time.” Most importantly, the VP stressed, “M learns from human behaviors.”11 The corporation’s machines would be trained on surplus from Messenger’s 700 million daily users. Eventually, it was hoped, M’s operations would be fully automated and would not need human trainers.
By 2017, Facebook had scaled back its machine intelligence ambitions and focused its personal assistant on the core mission: commerce. “The team in there now is finding ways to activate commercial intent inside Messenger,” a Facebook executive reported.12 The idea is to “prioritize commerce-driven experiences” and design new ways for users to “quickly buy things” without the tedium of entering credit card information, flipping pages, or opening applications. Pop-up buttons appear during your conversations with friends whenever the system detects a possible “commercial intention.” Just tap to order, buy, or book, and let the system do the rest.13
In this way the “personal digital assistant” is revealed as a market avatar, another Trojan horse in which the determination to render and monetize you
r life is secreted under the veil of “assistance” and embellished with the poetry of “personalization.” Its friendly recommendations, advice, and eagerness to act on your behalf barely conceal an aggressive new market cosmos hovering over any and every aspect of your daily life. It may be composed of restaurants, banks, plumbers, merchants, ticket sellers, airlines, and a limitless queue of possible strangers summoned by their interests in your behavior: now, soon, and later. They are standing by to cash in on your walk to work, your conversation with your teenager, or your aging running shoes. A digital assistant may derive its character from your inclinations and preferences, but it will be skewed and disfigured in unknown measure by the hidden market methods and contests that it conceals.
Google joined other tech companies determined to establish “conversation” as the medium through which humans engage with the apparatus. In time, the obsession with voice may be surpassed or joined by others so that merely thinking a thought or waving a finger can translate into and initiate action. For now, there are compelling reasons for the race to the spoken word. The first is obvious: reliable voice recognition can translate a sprawling landscape of service interactions into low-cost automated processes of theoretically unlimited scale and scope, a fact that has been noted by labor economists for some time.14 The competitive race between a new crop of “personal digital assistants” is best understood from this point of view. The voice that rises to dominance, the One Voice, will be the colossus of behavioral surplus pipelines with a potentially insurmountable competitive advantage in its ability to corner and kidnap the dominant share of human experience.
“Conversation” stands alone in its promise to dominate raw-material supply, and the rewards to the One Voice would be astronomical. Casual talk helps to blur the boundaries between “it”—the apparatus saturated with commercial agents—and us. In conversation we imagine friendship. The more we fancy the apparatus as our confidante, nanny, governess, and support system—a disembodied, pervasive “Mrs. Doubtfire” for each person—the more experience we allow it to render, and the richer its supply operations grow. Communication is the first human joy, and a conversational interface is prized for the frictionless ease in which a mere utterance can trigger action, especially market action: “Let there be light.” “Let there be new running shoes.” What could be dreamier than to speak and have it be so? An Amazon senior vice president comments on the company’s voice-activated home devices: “The nice thing about the Amazon device business is that when we sell a device, generally people buy more blue jeans. And little black dresses. And shoes. And so that’s good.” “Voice shopping,” he concludes, is good for business and good for predicting business.15
In a conversation addressed to a digital thing, as opposed to a conversation in a shop, words can occur on the fly with less friction and effort; less inhibition, fretting, and comparing; less concern about the limits of one’s bank account or where a product or service is sourced; less doubt and hesitation; less memory and remorse. The speaker feels herself at the center of a seamlessly flowing universe. The seams are all backstage, where the machines confront and conquer stubborn sources of friction such as distinct apps and entities; recalcitrant administrative service, distribution, payments, and delivery systems; and boundaries and borders that threaten the flows of desire and satisfaction. Spontaneous and fluid, universally burbling “conversation” turns the new personal digital assistant into a voice that sits between your life and the new markets for your life, between your experience and the auctioning of your experience: “a runtime, a new interface” that creates the sensation of mastery while, in fact, giving it away.
In this commercial dreamscape, words that were once conceived of as “behind closed doors” are eagerly rendered as surplus. These new supply operations convert your talk into behavior for surplus in two ways. The first derives from what you say, the second from how you say it. Smart-home devices such as Amazon’s Echo or Google Home render rivers of casual talk from which sophisticated content analyses produce enhanced predictions that “anticipate” your needs. Google used its 2016 developers conference to introduce its conversational reimagining of Google Now, rechristened “Assistant” and integrated across the company’s devices, services, tools, and applications. “We want users to have an ongoing, two-way dialogue with Google. We want to help you get things done in your real world and we want to do it for you,” explained Google CEO Sundar Pichai. “For example, you can be in front of this structure in Chicago and ask Google, ‘Who Designed This?’ You don’t need to say ‘the bean’ or ‘the cloud gate.’ We understand your context and we answer that the designer is Anish Kapoor.”16
Google’s Assistant arrived already integrated into Google’s new messaging app, Allo, where it can search for information, execute tasks, or even compose routine messages on your behalf. Most importantly, Assistant animates the firm’s home device, Google Home.17 The idea is that in time, the device (or its successor) will claim for rendition a theoretically limitless scope of animate and inanimate domestic activities: conversations, lightbulbs, queries, schedules, movement, travel planning, heating systems, purchases, home security, health concerns, music, communication functions, and more.
There was a time when you searched Google, but now Google searches you. Advertisements for Google Home feature loving families leading busy, intricate lives but visibly relieved to return home and fall into the arms of this omniscient, efficient caretaker. This second-modernity dream come true extracts an unusually high tax for its promise of a more effective life. For each user to have his or her own individual Google, as Pichai envisions, Google must have each individual.18
The caretaker’s ability to effectively serve you depends entirely upon the degree to which your life is rendered, knowingly or unknowingly, to its ministrations. The breadth and depth of rendered life correspond to the scale of market action that can be triggered and mediated by Assistant. There are differences among the various incarnations of “personalization” and “assistance” offered by the tech giants, but these are trivial compared with the collective urge toward total knowledge—about your inner states, real-world context, and specific daily life activities—all in the service of successfully training the machines that they might better target market operations to each moment of life.
All the potential market action associated with what you say depends upon voice activation, recognition, and response. These, in turn, are the product of highly sophisticated machine systems trained on vast global stockpiles of spoken words. The more structural insights the machines glean from spoken surplus, the more commerce flows from its content. This means that the value of what you say cannot be realized without machines that can learn from precisely how you say it. This form of surplus derives from the structure of your speech: vocabulary, pronunciation, intonation, cadence, inflection, dialect.
The competition for supplies of talk turns your phrases into this second form of surplus as companies determined to develop and perfect voice capabilities scour the world for speech. “Amazon, Apple, Microsoft and China’s Baidu have embarked on a worldwide hunt for terabytes of human speech,” reports Bloomberg Businessweek. “Microsoft has set up mock apartments in cities around the globe to record volunteers speaking in a home setting.” The tech firms capture flows of talk from their smart devices and phones as they record and retain your words. Chinese search firm Baidu collects speech in every dialect: “Then they take all that data and use it to teach their computers how to parse, understand, and respond to commands and queries.”19
Pieces of your talk are regularly farmed out in bulk to third-party firms that conduct “audio review processes” in which virtual scorers, tasked to evaluate the degree of match between the machine’s text and the original chunk of human speech, review audio recordings retained from smartphones, messaging apps, and digital assistants. Companies such as Amazon, Google, and Microsoft use these audio analyses to improve the algorithms of their voice systems. The tech compani
es insist that such recordings are anonymous, nothing more than voices without identities. “Partners do not have the ability to tie the voice samples back to specific people,” one Microsoft executive asserted. But one journalist who signed on to a virtual job as an audio recording analyst concluded just the opposite, as she listened to recordings full of pathos, intimacy, and easily identifiable personal information:
Within the recordings themselves, users willingly surrender personal information—information that is especially valuable in these review processes because they are so specific. Uncommon names, difficult-to-pronounce cities and towns, hyperlocal oddities.… I heard people share their full names to initiate a call or offer up location-sensitive information while scheduling a doctor’s appointment… the recordings capture people saying things they’d never want heard, regardless of anonymity.… There isn’t much to keep people who are listening to these recordings from sharing them.20
There is also substantial capital investment directed at talk, and Samsung’s Smart TV illustrates some of this behind-the-scenes action. Business forecasts routinely predict strong growth in the market for internet-enabled appliances, and Samsung is among a small group of market leaders. Its appliances use the Android operating system platform, and early on the firm established alliances with both the Alphabet/Google subsidiary Nest and with Cisco. “Our first mission is to bring your home to your connected life,” a top executive explained in 2014.21 In 2015 privacy advocates discovered that the corporation’s smart TVs were actually too smart, recording everything said in the vicinity of the TV—please pass the salt; we’re out of laundry detergent; I’m pregnant; let’s buy a new car; we’re going to the movies now; I have a rare disease; she wants a divorce; he needs a new lunch box; do you love me?—and sending all that talk to be transcribed by another market leader in voice-recognition systems, Nuance Communications.22
The Age of Surveillance Capitalism Page 32