Abstract
The more reliant we are on digital technologies, the more important the traces we leave behind become. While the device we type on, the browser we search with, and the number of times we reload Facebook may appear trivial, together they can help companies decide which products to hawk to us and, more unnervingly, they can help governments decide whether we are citizens or terrorists.
John Cheney-Lippold explores some of the more worrisome things that can occur when companies and governments stop thinking of people as humans and start thinking them only as valuable datasets. He argues that the misrecognitions and “mistakes” that may result from algorithmic processing are not errors but rather reconfigurations with material consequences, “freshly minted algorithmic truth that cares little about being authentic but cares a lot about being an effective metric for classification” (p. 16). Recording our lives into data necessarily transforms how others see us and how we think of ourselves.
We Are Data is part of a rapidly growing body of algorithmic culture scholarship from scholars, including Ted Striphas, Ed Finn, and Safiya Umoja Noble. What sets this work apart is its uncovering of many fascinating and often bizarre examples that stretch beyond the US border and across the cultural and political spectrum. Together, they illustrate the myriad ways that algorithms inflect our lives and shape our identities. In analyzing these examples, Cheney-Lippold draws on the long canon of feminist, queer, and critical theory regarding the relationship between technology and subjectivity. However, the many theoretical discussions throughout the book engage so many concepts that they often feel perfunctory and divorced from the chapters’ main arguments.
The book begins with a short preface about how Google has tried to automate new privacy regulations in the EU. Cheney-Lippold helpfully points out here—and throughout much of the rest of his book—how such algorithmic automation necessarily frames “how we come to understand ourselves and our place in the world” (p. 11). Following from the preface, the introduction explores how “datafication” has become not only lucrative to corporations, but also a tool of government surveillance and control that, intentionally or not, tends to replicate racist and sexist attitudes. He uses several examples, including the case of HP’s facial recognition software which could not identify Black people, to vividly illustrate how the question of what counts (or can be quantified) makes certain subjects invisible and can lead to a wide variety of discriminatory practices.
Cheney-Lippold goes on to explain the basics of how algorithms calculate and define identity. He provocatively compares the NSA’s reliance on metadata to identify terrorists to many other algorithmic categorization strategies, including how Google attempts to identify us based on what we do online and how a marimba playing robot tries to replicate the style of John Coltrane, Thelonius Monk, or anyone it is playing with. In my view, these algorithms appear to work quite differently (and, due to proprietary policies, there is no real way to know exactly how many of them work at all) and, hence, the comparisons between them are not always very convincing. Yet, Cheney-Lippold’s underlying point that companies and governments too often categorize us based solely on data, which can lead to anything from mildly annoying to potentially deadly results, is extremely compelling.
The later chapters include a wide variety of fascinating and esoteric examples that illustrate how algorithms affect “who we are, who we can be, and what can be uttered and understood about both” (p. 92). Numerous and wide-ranging examples include discussions of how Wii Sports calculates your “age”; how Google Flu failed to locate flu outbreaks; how UC Berkeley researchers tried to calculate the average or “essential” Santa; and how the US government created ZunZeo, the “Cuban Twitter,” in order to foment a revolution. While Cheney-Lippold importantly points out how all of these programs together illustrate that algorithmic thinking defines identity in very fluid ways, often far more attention is paid to their future potential rather than their current limitations. For instance, he offers a significant discussion of the harm of relying solely on algorithms to make decisions. In the case of the NSA’s targeting of terrorists via metadata this could be deadly. More often—as in the case of OkCupid’s misleading profiles or Google’s haphazard translation of certain common words into racial expletives—it is obnoxious and offensive. Rather than go deeply into any one example, Cheney-Lippold chooses to survey many. While this has certain benefits, it also at times replicates the same algorithmic logic, with its inherent problems of simplification and essentialism that he critiques.
Returning to the concerns brought up in the preface, Cheney-Lippold then considers how algorithms threaten the very ideal of privacy. While he critiques how privacy laws are often implemented, he fears that without it, we would “lack an integrity of the self” (p. 206). His primary example here is the emergency call of Mark Hemmings, who died primarily because the dispatcher’s algorithmically generated responses did not recognize his cries of pain to be an adequate sign of a life-threatening emergency. While it is not clear how Hemming’s death is related to privacy, it is a startling illustration of the real danger that treating people simply as a collection of data points and unthinkingly trusting algorithms can bring. With various other examples of privacy breaches and attempts to combat them, Cheney-Lippold calls for a revaluing of privacy in relation to our online data and illustrates the complexity of algorithmic surveillance and the lack of an obvious solution to this dilemma.
The conclusion reframes the book’s examples in relationship to techno-utopian discourses. While technologists tend to hope that “objective” and unhuman (or inhumane) algorithms will rid us of discrimination and inequality, instead they appear to automate and intensify such practices. The enormous value of data and the terribly lopsided power dynamics between users, companies, and governments guarantee that while our privacy continues to dissipate, companies like Google and Facebook become uncontrollable. Cheney-Lippold ominously ends by stating that our data selves are essentially “mathematical interpretation that makes you useful . . . for them” (p. 216).
