Abstract
Digital technologies claim to offer both a dazzling promise and a humbling nonpresence in users’ lives. For instance, while consumers use machines that are able to learn their voices and provide immediate services, they are also largely unaware of how and when these machines listen. Other users have immediate access to a seemingly unquantifiable amount of information online through search engines without having to worry much about the presence of algorithms that collate and present such information. YouTube’s autoplay function, for example, provides users with the ability to forefront their interests in a grassroots sort of way, all while searching and discovering fresh content from others, again without explicit and readable information about how such content appears on a user’s suggested page. In Algorithms of Oppression, Safiya Noble treks these surreptitious undercurrents of machine learning through Google’s search functions and argues that marginalized identities—and Black girls and women, in particular—are co-opted and grossly misrepresented through Google’s commercial model and algorithmic proprietary design. Noble offers an alternative, noncommercial model to think about information gathering online in ways that attend to societal inequities and oppressions.
Algorithms of Oppression takes an information science, critical race, and Black feminist perspective to offer practical ways to reimagine the corporate logic of search technologies due to their “willful neglect or a profit imperative that makes money from racism and sexism” (p. 5). Corporations like Google, Noble argues, operate from a “public good” business model in which users are led to believe that the information suggested by Google’s algorithms are both objective and in the interest of the public's advancement of knowledg. From this basis, Google’s search results become society’s ‘common sense.' Noble begins by dispelling such notions of objectivity regarding algorithms by pointing out, among other things, the inevitable human components that go into building these algorithms in the first place. The crux of her argument examines the corporate interests of Google as its search results necessarily bias information in favor of the company’s profitability and market dominance. Such commercialism drives the representative identity of Black women and girls online by making racist and sexist content profitable. Black women and girls are disproportionately, then, represented in stereotypical or pornographic ways, which subsequently affects how these users are targeted both ofline and online. Such is the territory, Noble argues, that spawns from the unregulated, unexamined territory of an ostensibly objective powerhouse like Google.
Chapter 2 “Searching for Black Girls” introduces the concept of technological racialization as a particular kind of algorithmic oppression that benefits some groups and marginalizes others. Though it operates in particular ways online, this oppression is of course rooted in historical and social processes of disenfranchisement. Noble explores the importance of search here by exploring online representation as power. She argues that dominant groups classify and organize representation of marginalized groups all while making such classification ‘common sense' to their users. Noble’s intervention is her incisive analysis of Black girls’ and women’s representation online as connected to the larger commercial structure that relies on such oppression. In this privatized digital realm, Black girls and women become a commodity within Google’s information system, which continues the legacy of white racial patriarchy at the expense of marginalized peoples. Chapter 3 “Searching for People and Communities” traces such online representation and connects it to off-line occurrences. For example, Noble analyzes the heinous 2015 case of Dylann Roof who opened fire and killed nine African-American worshippers at Emanuel African Methodist Episcopal Church in Charleston, South Carolina. She explores the actual consequences of search by revealing Roof’s Google searches, which allegedly included “black on White crime” in context of the Trayvon Martin police shooting in 2013. The results that Roof allegedly received largely came from sites like the Council of Conservative Citizens rather than statistics from, for example the FBI, on actual intraracial homicides. By conducting her own search on “black on White crimes” as well as “black girls” over a particular time period and logged out of personalized platforms, Noble found that much of the search results promote an incendiary amount of anti-Black hatred. Her larger argument here is about search engines as they are often utilized to “fact check” news media reports. However, there are serious concerns when these search engines, deemed as producing “facts,” return fascist, anti-Black, anti-Jewish, and white supremacist content that further feeds some users’ hatred. Noble continues that through results, search engines oversimplify complex phenomena, such as intraracial crime, which denies users the ability to critically engage with information. Here, Noble is clear about regulations that must be put in place regarding corporate entities such as Google who “compromise our ability to engage with complicated ideas” (p. 118).
In Chapter 4, “Searching for Protections from Search Engines,” Noble explores more fully the idea of regulatory digital systems. She points to countries, such as those in the European Union, who have implemented laws like the “right to forget and be forgotten,” as examples in which the government recognizes the importance of search as powerful records of individual and group representations. Noble argues that the right to be forgotten counters ideologies in the United States, which rely on digital corporations’ right to control information. Marginalized users in this process have little to no control over the ways that they are represented and remembered online. In this way, Google not only controls the records of our lives, but it owns our identity markers. Noble also critiques the relationship here between Google and the U.S. government, as the latter has “outsourced its data collection and unconstitutional privacy invasions to the company” (p. 129). The consequences and control of Google’s search then is far reaching and deeply entrenched in the U.S.
Chapter 5 “The Future of Knowledge in the Public” offers an alternative conceptualization of search that reprioritizes the interests of corporate entities for marginalized users. What’s at stake here is the categorization of knowledge and a look at who controls and has access to information. Noble cites existing search engines that are already doing the work of prioritizing particular communities, such as BlackWebPortal, which spotlights information pertinent for Black users, and Jewogle, a site dedicated to presenting the accomplishments of Jewish people. She then introduces a public search alternative that acknowledges the impact of information control online and the effects on marginalized bodies. This alternative would ideally, she says, work closely with public-interest journalism and librarianship “to ensure that the public has access to the highest quality information available” (p. 152).
Safiya Noble ends by theorizing and advocating for a Black Feminist Technology Studies that would allow for an epistemological approach to studying racialized and gendered identities within media studies. Research using this framework would prioritize intersectional identities and offer counternarratives about Black people and technology. Noble’s intervention through Algorithms of Oppression is thus both theoretically useful for future researchers as well as practical for policy makers and regulators. The book is a needed critique of large digital structures that permeate millions of users’ lives and decisions every day. She centers Black girls and women in a way that most industry work simply does not. While Noble makes these integral interventions, the book leaves room for others to pick up on methodological discussions about the precise ways that we study algorithms and marginalization online. That is, given the ‘moving target' and proprietary nature of algorithms, how do we begin to think about technologies that similarly track and hold corporate bias accountable? Are user-based algorithms possible in the direction of continuing to reveal, on a broad yet communitarian scale, the impacts of algorithms that Noble carefully charts? Or are there necessary drawbacks when quantifiably critiquing these systems of power? Noble’s book has laid the groundwork for these and other questions into the relationship between digital technologies and marginalized identities as well as the inevitable surrounding issues of power, knowledge, control, and resistance.