Behind the Screen: Content Moderation in the Shadows of Social Media by Sarah T. Roberts (review)

"Things aren't great, Internet. Actually, scratch that: they're awful."¹ This quote from an open letter to the Internet published by Wired succinctly summarizes the current state of online affairs, including broken expectations and unfulfilled promises once pegged upon the development of Internet-connected information and communication technologies. That the Internet brought many wonderful services is indisputable. What causes incessant alarm is the dark side of it, namely, "the pornographic, the obscene, the violent, the illegal, the abusive, and the hateful" that circulates on social media platforms.²

Launched to further correct "a collective myopia" (61) regarding technology, Sarah T. Roberts's study inquires into the practices of what she calls "commercial content moderation" (CCM). Her exploration adopts a very strict definition of CCM as "a job, a function, and an industrial practice that exists only in this context and could only ever exist in it" (15). By separating online content moderation from the practices of censorship, the author insulates her study from the political powerlines that it inevitably crosses and proceeds to find out who does content screening for tech companies, what their labor looks and feels like, and what it entails.

From the author's own experience with digital technology and her observations that the digital hides—or, worse, erases—human traces of its making and maintenance, the narrative moves to present three groups of content moderators: in-house moderators employed by MegaTech (a pseudonym) in Northern California; "boutique" moderators (41) hired to manage the clients' online presence (represented by a Canadian expat in Mexico and a former contractor for a digital news company); and workers in business process outsourcing services in Manila, the Philippines. Freelancers and those viewers who voluntarily flag offensive and inappropriate content on "microlabor websites" (46) are mentioned only briefly, since the main focus of the study is on "professional people who are paid to screen content" on social media platforms (1). The interviews with content moderators unearth several common themes across locations where the task of moderation is performed: precarious work, pervasive stress and burnout, job insecurity, disparity in status and treatment—all triangulating on a conclusion that CCM is invisible by design and that content moderators are invisible by association and hence easily disposable as workers, since their one-year contract allows one renewal only (82).

Extended quotations from the interviews are the strongest element of the book, vividly detailing the content moderation experiences of moderators that the platforms purposefully keep hidden from the public eye. Together, these corroborate evidence that in its "rote, repetitive, and endless work tasks" (92) content moderation frequently resembles labor on a factory's assembly line (92). Still, out of the many dramatic images and takeaways that this book offers, two stand out. One is that the romantic notion of the Internet as a free-speech zone is largely blind to the reality of the Internet as "a site of control, surveillance, intervention and circulation of information as a commodity" (6). And the other is that without such content moderation, the Internet "would be largely unusable for most people," as it would likely turn into a "cesspool" of aggression and negativity (153); consequently, the companies will continuously need human screeners and "data janitors."³

With insights from six interviewees, the book presents an excellently executed act of raising awareness about the work done behind the screen. However, because the analysis is light and accessible to the general audience, it misses a chance to build an argument in support of advocacy for change in the conditions under which much digital information labor is performed or for a broader public discussion of policies, practices, and patterns of the gig economy. That such a discussion is urgently needed is evident in the growing number of studies that reveal the legacy of inequality and discrimination in many technologically restructured service industries...