The Age of Surveillance Capitalism

by Shoshana Zuboff

  Google’s mission was to “organize and make accessible the world’s information,” and by late 2001 the intelligence community established “information dominance” in the public’s house, quickly institutionalizing it in hundreds of billions of dollars’ worth of state-sponsored global technology infrastructure, personnel, and practice. The contours of a new interdependency between public and private agents of information dominance began to emerge, one that is best understood through the lens of what the sociologist Max Weber once called an “elective affinity” born of a mutual magnetism that originates in shared meanings, interests, and reciprocities.70

  The elective affinity between public intelligence agencies and the fledgling surveillance capitalist Google blossomed in the heat of emergency to produce a unique historical deformity: surveillance exceptionalism. The 9/11 attacks transformed the government’s interest in Google, as practices that just hours earlier were careening toward legislative action were quickly recast as mission-critical necessities. Both institutions craved certainty and were determined to fulfill that craving in their respective domains at any price. These elective affinities sustained surveillance exceptionalism and contributed to the fertile habitat in which the surveillance capitalism mutation would be nurtured to prosperity.

  The elective affinity between public and private missions was evident as early as 2002, when former NSA Chief Admiral John Poindexter proposed his Total Information Awareness (TIA) program with a vision that reads like an early guide to the foundational mechanisms of behavioral surplus capture and analysis:

  If terrorist organizations are going to plan and execute attacks against the United States, their people must engage in transactions and they will leave signatures in this information space.… We must be able to pick this signal out of the noise… the relevant information extracted from this data must be made available in large-scale repositories with enhanced semantic content for analysis to accomplish this task.71

  As CIA Director George Tenet had declared in 1997, “The CIA needs to swim in the Valley,” referring to the need to master the new technologies flowing from Silicon Valley.72 In 1999 it opened a CIA-funded venture firm in the valley, In-Q-Tel, as a conduit for cutting-edge technologies. The operation was meant to be an agency experiment, but after 9/11 it became a critical source of new capabilities and relationships, including with Google. As Silicon Valley’s Mercury News reported, “There’s a new urgency with the CIA to find technology that makes sense of all the unstructured data floating around on the internet and elsewhere. The agency can’t train analysts quickly enough.” In-Q-Tel’s CEO described the government’s agencies as “scrambling” and noted that “we’re in a state of hyperactivity now.”73

  Surveillance exceptionalism thrived in that hyperactivity. Poindexter’s Total Information Awareness program did not obtain congressional support, but an analysis in the MIT Technology Review showed that many of the TIA initiatives were quietly reassigned to the Pentagon’s Advanced Research and Development Activity (ARDA), which in 2002 received $64 million to fund a research program in “novel intelligence from massive data.” In 2004 the US General Accounting Office surveyed 199 data-mining projects across dozens of federal agencies and more than 120 programs developed to collect and analyze personal data to predict individual behavior.74 The New York Times reported in 2006 that the intelligence agencies, backed by a $40 billion annual budget, regularly fielded secretive shopping expeditions to Silicon Valley in search of new data-mining and analysis technologies.75

  State security agencies sought ways to avail themselves of Google’s rapidly developing capabilities and simultaneously use Google to further develop, commercialize, and diffuse security and surveillance technologies with proven intelligence value. If TIA could not be fully developed and integrated in Washington, parts of the job could be delegated to Silicon Valley and its standout in information dominance: Google. By late summer 2003, Google was awarded a $2.07 million contract to outfit the agency with Google search technology. According to documents obtained by Consumer Watchdog under the Freedom of Information Act, the NSA paid Google for a “search appliance capable of searching 15 million documents in twenty-four languages.” Google extended its services for another year at no cost in April 2004.76

  In 2003 Google also began customizing its search engine under special contract with the CIA for its Intelink Management Office, “overseeing top-secret, secret and sensitive but unclassified intranets for CIA and other IC agencies.”77 Key agencies used Google systems to support an internal wiki called Intellipedia that allowed agents to share information across organizations as quickly as it was vacuumed up by the new systems.78 In 2004 Google acquired Keyhole, a satellite mapping company founded by John Hanke, whose key venture backer was the CIA venture firm, In-Q-Tel. Keyhole would become the backbone for Google Earth, and Hanke would go on to lead Google Maps, including the controversial Street View Project. In 2009 Google Ventures and In-Q-Tel both invested in a Boston-based startup, Recorded Future, that monitors every aspect of the web in real time in order to predict future events. Wired reported that it was the first time the CIA-backed venture firm and Google had funded the same startup and that both firms had seats on Recorded Future’s board of directors.79

  In the decade that followed 9/11, surveillance exceptionalism was also expressed in the flattery of imitation, as the NSA tried to become more like Google, emulating and internalizing Google’s capabilities in a variety of domains. In 2006 General Keith Alexander outlined his vision for a search tool called ICREACH that “would allow unprecedented volumes of… metadata to be shared and analyzed across the many agencies in the Intelligence Community.” By late 2007, the program was piloted, boosting the number of communications events it shared from 50 billion to more than 850 billion. The system was designed with a “Google-like” search interface that enabled analysts to run searches against meta-data “selectors” and to extract vital behavioral surplus for analyses that could reveal “social networks,” “patterns of life,” and “habits,” and in general “predict future behavior.”80 In 2007 two NSA analysts wrote an internal training manual on how to find information on the internet. It expressed the agency’s keen interest in all things Google with a detailed chapter devoted to a deconstruction of Google Search and the Google “hacks” that can uncover information not intended for public distribution.81

  That year, the elective affinities that infused the intelligence community’s interest in Google were also highlighted when Google research director and AI expert Peter Norvig presented at a Pentagon Highlands Forum meeting: an exclusive networking event where military and intelligence officials commune with members of the high-tech industry, elected officials, elite academics, top corporate executives, and defense contractors. In 2001 the forum’s director, Richard O’Neill, described its work to a Harvard audience as “an idea engine, so the ideas that emerge from meetings are available for use by decision makers as well as by people from the think tanks.”82 It was to be a bridge between the government and commercial leaders, especially in Silicon Valley.83 According to one highly detailed account by investigative journalist Nafeez Ahmed and cited by legal scholar Mary Anne Franks, the forum was both a support system and an incubator of Google’s growth, as well as a connecting and convening force for the Pentagon, intelligence agencies, and the young company: “The US intelligence community’s incubation of Google from inception occurred through a combination of direct sponsorship and informal networks of financial influence, themselves closely aligned with Pentagon interests.”84 Another legal scholar described the “collaboration” between Google and the intelligence community, especially the NSA, as “unprecedented.”85

  During these years, scholars noted the growing interdependencies between the intelligence agencies, resentful of constitutional constraints on their prerogatives, and the Silicon Valley firms.86 The agencies craved the lawlessness that a firm such as Google enjoyed. In his 2008 essay “The Constitution in the National
Surveillance State,” law professor Jack Balkin observed that the Constitution inhibits government actors from high-velocity pursuit of their surveillance agenda, and this creates incentives for the government “to rely on private enterprise to collect and generate information for it.”87 Balkin noted that the Supreme Court has imposed few privacy restrictions on business records and information that people give to third parties. E-mail is typically held in private servers, making its protection “limited if not nonexistent.” This absence of law made private companies attractive partners for government actors bound to democratic constraints.

  The government’s need to evade constitutional oversight, argues legal scholar Jon Michaels, leads to secret public-private intelligence collaborations that tend to be “orchestrated around handshakes rather than legal formalities, such as search warrants, and may be arranged this way to evade oversight and, at times, to defy the law.”88 He observed that intelligence agencies are irresistibly drawn to “and in some respects dependent upon” firms’ privately held data resources.89

  Both scholars’ observations were confirmed in 2010, when former NSA Director Mike McConnell offered another glimpse into the elective affinities between Google and the intelligence community. Writing in the Washington Post, McConnell made clear that Google’s surveillance-based operations in data capture, extraction, and analysis were both taken for granted and coveted. Here the boundaries of private and public melt in the intense heat of new threats and their high-velocity demands that must be met in “milliseconds.” In McConnell’s future there is one “seamless” surveillance empire in which the requirements of self-preservation leave no opportunity for the amenities of democracy, with its time-wasting practices of due process, evidence, warrants, and law. As McConnell insisted,

  An effective partnership with the private sector must be formed so information can move quickly back and forth from public to private and classified to unclassified… to protect the nation’s critical infrastructure. Recent reports of possible partnership between Google and the government point to the kind of joint efforts—and shared challenges—that we are likely to see in the future… such arrangements will muddy the waters between the traditional roles of the government and the private sector.… Cyberspace knows no borders, and our defensive efforts must be similarly seamless.90

  In the final months of the Obama administration, then Secretary of Defense Ash Carter toured Silicon Valley, where he announced a new Defense Innovation Advisory Board, meant to formalize a channel between the tech executives and the DOD. Carter appointed Schmidt to the new board and tasked him with selecting its members. As Wired concluded, “The government needs Silicon Valley more than ever as it seeks to defend from security threats in cyberspace.”91 These facts are amply illustrated in a comprehensive treatment of “bulk collection” by an international group of scholars and edited by Indiana University’s Fred Cate and Berkeley’s James Dempsey. Cate and Dempsey note the “expansive aggregation” of personal data in the hands of private companies: “Governments understandably want access to this data.… Essentially every government in the world claims the power to compel disclosure of this data by the companies that hold it.”92 Had it not been for surveillance exceptionalism, it is possible that these data would not even exist, at least not in their current volume and detail.

  Surveillance exceptionalism helped to shape the evolutionary course of information capitalism by creating an environment in which Google’s budding surveillance practices were coveted rather than contested. Once again, history offers us no control groups, and we cannot know with certainty whether information capitalism might have developed in a different direction had it not been for the sudden new interest in surveillance capabilities. For now, it appears that one unanticipated consequence of this public-private “elective affinity” was that the fledgling practices of surveillance capitalism were allowed to root and grow with little regulatory or legislative challenge, emboldening Google’s young leaders to insist on lawlessness as a natural right and, in ways that are even more opaque, emboldening the state to grant them that freedom.

  Powerful elective affinities favored the acquisition of certainty at any price, and part of that price appears to have been the shelter of surveillance capitalism. In the fullness of time, historians will no doubt discover the specifics of these relationships and the ways in which Google’s discoveries in the capture and use of behavioral surplus were sheltered from scrutiny, at least in part, because of this new habitat of militarized demand.

  In the context of new military purpose, the digital capabilities that were aimed toward the advocacy-oriented values of the behavioral value reinvestment cycle flowed toward surveillance without impediment. Surveillance assets thrived without risk of sanction and attracted surveillance capital. Revenue followed. The situation recalls the auto, steel, and machine tool industries at mid-century, when military orders kept plants operating at full capacity. In the end, however, this turned out to be more curse than blessing. Military demand distorted and suppressed the innovation process and drove a wedge between these industries and their civilian customers, leaving them vulnerable to foreign competitors in the globalizing markets of the late 1970s and early 1980s.93

  Similarly, under the conditions of surveillance exceptionalism, Google’s leaders were not compelled to undertake the arduous and risky work of inventing an exchange-based advocacy-oriented market form when the surveillance model was so lucrative. Why risk experimentation with more-organic paths to monetization when surveillance and extraction operations were safe from law and hugely profitable? Eventually, it wasn’t just Google asking these questions; every other internet business faced the same choices. Once surveillance revenues set the bar for venture capitalists and Wall Street analysts, it became that much easier for internet companies to go with the flow. Then it became onerous not to.

  V. Fortifications

  Why is it that so many years after the events that triggered the mania for information dominance, surveillance capitalism still runs with relatively little impediment, especially in the US? The intervening years have seen the proliferation of thousands of institutional facts that normalized surveillance capitalism’s practices and made them appear necessary and inevitable: the discovery of behavioral surplus and the massive accumulations of capital and material that followed, the proliferation of devices and services, the integration of data flows, and the institutionalization of futures markets in human behavior.

  This does not mean that we should succumb to the natural fallacy and interpret this flourishing as a signal of surveillance capitalism’s inherent worthiness or inevitability. In the coming chapters we will uncover many additional factors that have contributed to this success, but here I want to focus on Google’s proactive efforts to build fortifications around its supply chains in order to protect surplus flows from challenge.

  Although many elements of this fortification strategy have been well publicized, their importance for our story lies in the fact that each is one aspect of a multipronged effort that deflects scrutiny from core operations in order to maintain the flow of free, unregulated behavioral surplus. Fortifications have been erected in four key arenas to protect Google, and eventually other surveillance capitalists, from political interference and critique: (1) the demonstration of Google’s unique capabilities as a source of competitive advantage in electoral politics; (2) a deliberate blurring of public and private interests through relationships and aggressive lobbying activities; (3) a revolving door of personnel who migrated between Google and the Obama administration, united by elective affinities during Google’s crucial growth years of 2009–2016; and (4) Google’s intentional campaign of influence over academic work and the larger cultural conversation so vital to policy formation, public opinion, and political perception. The results of these four arenas of defense contribute to an understanding of how surveillance capitalism’s facts came to stand and why they continue to thrive.

  First, Google demonstrated th
at the same predictive knowledge derived from behavioral surplus that had made the surveillance capitalists wealthy could also help candidates win elections. To make the point, Google was ready to apply its magic to the red-hot core of twenty-first-century campaigning, beginning with the 2008 Obama presidential campaign. Schmidt had a leading role in organizing teams and guiding the implementation of cutting-edge data strategies that would eclipse the traditional political arts with the science of behavioral prediction.94 Indeed, “At Obama’s Chicago headquarters… they remodeled the electorate in every battleground state each weekend… field staff could see the events’ impact on the projected behaviors and beliefs of every voter nationwide.”95

  Research by media scholars Daniel Kreiss and Philip Howard indicates that the 2008 Obama campaign compiled significant data on more than 250 million Americans, including “a vast array of online behavioral and relational data collected from use of the campaign’s web site and third-party social media sites such as Facebook.…”96 Journalist Sasha Issenberg, who documented these developments in his book The Victory Lab, quotes one of Obama’s 2008 political consultants who likened predictive modeling to the tools of a fortune-teller: “We knew who… people were going to vote for before they decided.”97

  Obama used his proximity to Schmidt to cement his own identity as the innovation candidate poised to disrupt business as usual in Washington.98 Once elected, Schmidt joined the Transition Economic Advisory Board and appeared next to Obama at his first postelection press conference.99 According to Politico, “The image alone of Schmidt standing elbow-to-elbow with Obama’s top economic thinkers was enough to send shivers up the spine of Google’s competitors. ‘This terrifies Microsoft,’ said a Democratic lobbyist familiar with the industry. ‘There’s a reason why people are scared to death of Google.’”100