What Do We Critique When We Critique Technology?
2023; Duke University Press; Volume: 95; Issue: 2 Linguagem: Inglês
10.1215/00029831-10575091
ISSN1527-2117
Autores Tópico(s)Modern American Literature Studies
ResumoThinking about the state of technology today necessarily means thinking about a number of interrelated but distinct entities. Considering the nuts and bolts of a news story in which, say, some corporate machine vision technology was found to be racially discriminatory can often mean having to study business practices, data sciences, specific suites of tools that can lay a claim to the moniker of AI, assemblages of hardware and software, platform infrastructures with machines slotted away in hot data-center basements in tax havens, human-computer interactions and perceptions, and academic/industry discourses within any of the aforementioned, not to mention the geopolitical and historical situation of it all, which may further call into question where, say, "American" literature can uniquely intersect with technologies splayed awkwardly across, and not always along, the traditional geopolitical and cultural fault lines. In such a scenario, the flag of "Critical AI and (American) Literature," by its very constitution, carries several sigils, including those of big data and literature and of computational culture and literature, as well as American studies and global technological sovereignties. Focused on the more critical end of these studies, this review brings together three new multiauthored books to ask what we critique when we critique technology today.Scholars interested in literature and technology—usually found in disciplines and departments such as languages and literatures, cultural studies, science studies, and media studies—have long been producing pathbreaking critical thought about various sociotechnical phenomena. From reading technologies themselves using literary critical methods—N. Katherine Hayles, Donna Haraway, Friedrich Kittler, Wendy Chun, Matthew Kirschenbaum, Rita Raley, Lisa Gitelman, and Alexander Galloway all come to mind here—to studying literary expressions of technological worlds (see, e.g., work by Fredric Jameson, Bruce Clark, Laura Otis, Steven Shaviro, Sherryl Vint, and Colin Milburn), literary criticism has been a bellwether of technology critique for several decades now. A brief look at such critique through the ages shows us the varied moods that orient studies of these technologies, with AI just being the latest in this series that once featured the internet, the personal computer, hypertext, cellphone, and metadata. Where there was once a utopian dream with the expansion of networks in the 1990s, or a reluctant acceptance that became a residual flicker of counterprogrammatic hope that technologies can be reappropriated by radical social forces in late 2010s, there is now, in critical work collected here, largely anger and disappointment. Every day, as news cycles tell tales of unchecked tech monopolies roughly intruding into our social, political, and psychic lives, and rarely for the good, these authors find themselves angry—really angry—about the state of our technologies and what they have wrought. On the one hand, such anger indexes our historical condition and informs our engagement with technology today. On the other hand, it forces us to ask what we are actually angry about, and what can be done instead.The primary example of this mood may be found in Your Computer Is on Fire, edited by Thomas S. Mullaney, Benjamin Peters, Mar Hicks, and Kavita Philip. This volume is a startlingly direct collection of essays that, for the most part, all do what they say; the overarching purpose of the volume is, in fact, a call to action that signals a diffused state of emergency in various corners of computational cultures (6). The three parts of the book—"Nothing Is Virtual," "This Is an Emergency," and "Where Will the Fire Spread?"—contain chapters that are thematically and methodologically varied but all united by their clear and accessible critiques that point out how inequalities and discriminations are enabled and exacerbated by technological systems today. To note a few, Nathan Ensmenger's "The Cloud Is a Factory," which uses Amazon as a case study for the infrastructural reinscription of older techniques of capital used by Sears and Standard Oil, is an excellent breakdown of the material behind the supposedly virtual cloud (29); Ben Peters's "A Network Is Not a Network" accounts for the role of institutional behavior in the constitution of large-scale networks (71); Mar Hicks, through an analysis of gender discrimination in mid-twentieth-century technology labor sector, claims that "Sexism Is a Feature, Not a Bug" in tech economies and communities (135); Safiya Umoja Noble tells us that our robots aren't neutral (199); Janet Abbate takes on tech sector's consideration of coders using the pipeline model—the discourse that encourages getting more women and minorities to learn coding earlier and faster so as to facilitate a smoother and more expansive flow of more diverse labor into the infamously white and masculine technology sector—to show how "Coding Is Not Empowerment" (253); Ben Allen magisterially demonstrates how the same genre of technical hacks can be read as playful or criminal depending on power dynamics (273); and Paul Edwards studies platforms, which he calls fast infrastructures, as he takes up exemplars from South Africa and Kenya to suggest that these fleeting operational levers represent the next model of corporate infrastructural dominance (313). The volume, then, contains a series of related, but not necessarily coagulated, critiques of technology and its sociocultural conditions.Technoprecarious, collectively authored by the Precarity Lab (the contributing team comprises Cassius Adair, Iván Chaar López, Anna Watkins Fisher, Meryem Kamil, Cindy Lin, Silvia Lindtner, Lisa Nakamura, Cengiz Salman, Kalindi Vora, Jackie Wang, and McKenzie Wark) reads not like an edited collection but more like a short manifesto written by scholars with complementary orientations. The different sections of the book—among them "The Undergig," "The Widening Gyre of Precarity," "Automating Abandonment," "Fantasies of Ability," and"Dispossession by Surveillance"—come together to form a patchwork of commentaries, most well rooted in original cultural, sociotechnical, anthropological, and historical research, that all very playfully point out the exacerbation of precarity wrought by, with, and through digital technologies today. The titular technoprecarity, for the collective, is "the premature exposure to death and debility that working with or being subjected to digital technologies accelerates" (1). Technoprecarity here shows up in snippets that plug into work on surveillance, carceral systems, toxicity, and administrative failures, among other nodes of inquiry. The final two chapters feature a Haraway-esque hope for radical reappropriation listing the Detroit Digital Stewards Program, which features groups that help underprivileged communities gain access to technologies as tools of communication, and the use of open-source maps in Palestine as examples to be followed for practices of techno-oriented care (74–86). There is a sincere attempt here, not unlike the penultimate contribution in Your Computer Is on Fire, "How to Stop Worrying about Clean Signals and Start Loving the Noise" by Kavita Philip (363), to find a nugget or two of hope in the middle of the general condition of technoprecarity being described.Uncertain Archives: Critical Keywords for Big Data, edited by Nanna Bonde Thylstrup, Daniela Agostinho, Annie Ring, Catherine D'Ignazio, and Kristin Veel, is a six-hundred-page collection that features sixty-one keyword entries, altogether providing a Raymond Williams–style vocabulary for critical studies of data and AI. Considering big data as an uncertain archive—drawing from archival theory and critical data studies while thinking about the latent possibilities of aggregation as presented by big data today—the collection features short, punchy nuggets of wisdom that offer a polysemic understanding of the kinds of critical thought different disciplines can offer to studies of big data and AI at large. The entries vary widely in style, tone, content, and orientation. Overlapping questions of epistemologies ("Quantification" by Jacqueline Wernimont, "Ethics" by Louise Amoore, "Unpredictability" by Elena Esposito, "Remains" by Tonia Sutherland), alterity and discrimination ("Abuse" by Sarah Roberts, "(Mis)Gendering" by Os Keyes), power ("DNA" by Mél Hogan, "Instrumentality" by Luciana Parisi, "Organization" by Timon Beyes), aesthetics ("Demo" by Orit Halpern; a brilliant, poetic one on "Throbber" by Kristoffer Ørum; "Visualization" by Johanna Drucker), infrastructures ("Cooling" by Nicole Starosielski, "Supply Chain" by Miriam Posner, "Field" by Shannon Mattern), and socialities ("Values" by John S. Seberger and Geoffrey C. Bowker, "Proxies" by Wendy Hui Kyong Chun, Boaz Levin, and Vera Tollmann; "Self-Tracking" by Natasha Dow Schüll) all sit alongside questions near and dear to literary critical approaches ("Digital Humanities" by Roopika Risam, "File" by Craig Robertson, "Misreading" by Lisa Gitelman). In performing the immensely unenviable task of shepherding sixty-eight other scholars from across the world into one contained collection, the editors here provide a deliberately fragmented mise-en-scène of critical data studies as it unfolds across several corners of academia today. Juggling several different approaches, Uncertain Archives does not easily offer a shared through line. Nevertheless, it can be read as a collection trying to enumerate the various evaluative frameworks that can be applied to/in critical (big) data studies; most terms offered here—some of which are compressed versions based on the concepts outlined in the contributors' monographs and articles—can be taken as pedagogical scaffolds or starting points for a broader set of research inquiries. And big data here shows up as nebulous and tentacular, both in its contemporary material reach and in its analytic demand for methodological diversity.Before returning to talk about criticism and technology at scale again by considering what this cluster teaches us methodologically, let us first consider something like a state-of-the-field snapshot, briefly analyzing the who, what, and how of technology criticism presented by this sample set of edited collections, which is admittedly numerically small yet also somehow, by the virtue of including more than ninety authors from across the world, huge.It is in their conscious interdisciplinarity, tending toward a kind of postdisciplinarity, that all these three reviewed works are united in their orientation. Methodologically, they all occupy slightly different niches: Your Computer Is on Fire tries to advance one very broad but largely unified argument, Uncertain Archives presents a curated cacophony of positions, and Technoprecarious is a tight-knit manifesto that, quite uniquely and successfully, speaks in one voice. But in orientation, they are all joined at the hip, insofar as all these works bring together scholars of media studies, communication studies, digital studies, film studies, science and technology studies, data science, history and philosophy of science and technology, cultural studies, journalism, information and library sciences, American studies, critical race studies, postcolonial studies, gender and sexuality studies, organization studies, critical legal studies, critical geographies, digital humanities, and anthropology, along with literary critics. The fact that this list of disciplinary orientations seems overwhelming, despite several shared positions and investments of the many disciplines therein, is in fact one of the problems that this swell of collected work set out to address. In the special zone occupied by edited collections and coauthored work within academic circuits, these volumes (and others recently, including not only the very special issue you are perusing right now but also the 2022Critical Inquiry special issue on "Surplus Data" and the 2020 PMLA collection titled "Varieties of Digital Humanities") all mark a simultaneous expansion and contraction. On the one hand, this proliferation of edited collections showcases the best practices and results of bringing together different disciplinary research paradigms that such venues can perform, not to mention the new and exciting kinds of writing that coauthoring can wring out. On the other hand, the constant carving out of subdisciplinary niches (and the disillusionment from the same) in fact indexes the aftereffects of the shrinkage of academic positions per se, with every scholar herein wearing several hats and delving (or, perhaps more accurately, having to delve) into several distinct, if concomitant, disciplines.If one were to shift the focus from who is doing the study to how, one would find in some of this newer scholarship a desire to rethink the relationship between the scholar and the subject. Thinking with, instead of about, is well demonstrated across these collected works. For example, Technoprecarious, written by Precarity Lab, clearly starts out by staging its own existence as a form of laboratorial practice; the opening salvo states how [we] have adopted the "laboratory" (in our name and practice) to account for our highly ambivalent yet deeply entangled position in relation to ongoing attempts to upgrade and entrepreneurialize the humanities and scholarship and higher education broadly. The laboratory is a place of labor, but where labor is subordinated to the task of elaboration. In the lab, there are consistent procedures, forms of regularity that produce observable difference. The lab experiments—experiments that can be tested, verified, stabilized, and can become the prototypes for new forms of organization and governance. (8)From this position of being a laboratory (and rightly being very conflicted about its constitution as a laboratory), the collective then stages a worthy conversation around laboratories in history (of science, technology, and culture) and popular imagination. The text rejects the idea of the city as a laboratory (9), provides a proposition about laboratories of slavery as historical and conceptual precursors to logics that widen the gyre of precarity and uneven expansion of surveillance (34), and enables a critical consideration of Detroit as a laboratorial model, both for a public imagination that sees it as exemplifying the decline of American cities and for the activists whose efforts are taken up in the text to demonstrate some kind of resistance and response in face of late capitalist horrors (76). In other words, the book is an experiment produced by a laboratory that studies different kinds of laboratorial impulses critically but also generatively. Such thinking with, as the books collected here showcase, has scholars using the subject of inquiry not as a curiosity to be observed from a distance, not just thought about, but as something to be used and thought through until the subject in question animates other parts of one's inquiry. (See also Lindsay Thomas's contribution to this special issue.)There is also something to be said about what kind of epistemological work this interdisciplinarity does. Allow me to contextualize this aforementioned critical approach within the broader landscape of work around technology and culture. Humanistic studies of technology often provide, on the one hand, critiques of technological cultures themselves, and on the other, explorations of interrelations between technology and X, where X could be literature, media, or other sociocultural niches (for example, the whole domain of literature and science offers a set of methodologies for precisely this). Essentially, technocritical approaches, considered in this discussion by virtue of the books under review, are not afraid of critiquing technology at the level of technics, with criticality and technicality both in conversation with each other. Now, of course, the wedge between, say, cultural-critical and technocritical approaches isn't even a solidly distinct demarcation; plenty of work mentioned above ably flouts and flutters between these boundaries. However, one can still trace a faint distinction between work that stays far from the machinic cores and work that likes to stick close to the beating metallic heart of these systems. While the former might delve into the uses of technology—exemplified by, say, the aesthetic studies of electronic literature or the use of virtual reality in drama—the latter is likely to talk about the physical (or virtual) machines themselves. One is reminded here of the late technology theorist Leo Marx, whose complaint that technology is a hazardous concept lay atop the startlingly capacious denotations that the word consumed unto itself. The machines; the people who operated them; the systems that produced, sold, and maintained them; the infrastructural and politicoeconomic assemblages surrounding them were, per Marx (1997), rolled into one compact word sometime in the late nineteenth century in the United States: technology. In other words, the conceptual genesis of technology was a thoroughly American enterprise, with this nation-state offering the precise location for blurring economy, politics, and technics into one word. Technocritical approaches today mark a worthy (machinic) addition to the pantheon of approaches and orientations that break apart this word and that can be drawn on to study our world by providing a method of reading the nonhuman technologics for cultural, social, aesthetic, and political connotations. Here, in the variety of methodologies brought together by the books reviewed here, we can note a careful, if partial, fulfilment of the critical position exhorted by Marx; as noted below, scholars today are indeed studying technology as the aggregation of those capacities that Marx encountered in his concept genealogy. Within this matrix, technocriticality shows up as an approach that takes technology to include techniques and sociocultural logics while also not shying away from the machinic processes—demonstrated, for example, by Ben Allen's incisive reading of hack(ed/ing) code in Your Computer Is on Fire (273) and an excellent technopolitical breakdown of "Executing" by David Gauthier, Audrey Samson, Eric Snodgrass, Winnie Soon, and Magdalena Tyżlik-Carver (all members of the Critical Software Thing collective) in Uncertain Archives (209)—that used to be otherwise, sometimes unfortunately, black-boxed by earlier approaches.In other words, these works show how "the material component—technology narrowly conceived as a physical device—is merely one part of a complex social and institutional matrix" (Marx 1997: 979). While the presence of several different disciplinary orientations in scholarly collections such as the ones reviewed here resolves prima facie some problems of focal vagueness through an ever-increasing array of analytic subjects, handled with care and rigor, it also brings forth a new kind of lacuna: causal clarity. Today, the primary form of arguments—especially arguments that are able to exceed local disciplinary scholarship and move toward interdisciplinary or public writing—is often about locating problems within something presumed "neutral." Regardless of what in the technology matrix is being critiqued—"Facebook doesn't actually connect, and Google's algorithms are racist" (machines), "Silicon Valley has a diversity problem that it is actively not solving" (communities), and "gender discrimination permeates hierarchies in tech industry" (systems), to take a few examples—it is always, as Bruno Latour (2004: 240) stated in his screed "Why Has Critique Run Out of Steam?," about "debunk[ing] objects [that the critics] don't believe in by showing the productive and projective forces of people." Who, one may ask, needs to know that this or that technology is not neutral? Is it the layperson, the technoscientific disciplinarian, or perhaps the critic more than anyone else? Breaking the problem into constituent historically and geographically situated units, as Uncertain Archives does, we can roughly taxonomize technology critique itself today through five nondistinct kinds of objects of critique as handled by scholars who aim their interventions at three nondistinct kinds of audiences (see table 1). The various components of technology pointed out by theorists and historians of technology—physical machines, technological culture and art, the people behind and in front of these machines, systems and structures that enable and maintain the machines, and the techniques and habits that go through/with the machines—all find different (sub)disciplinary configurations studying and critiquing them not only in these reviewed texts but also in other critical scholarship generally. The work itself can easily flow between hyperspecific scholarship that targets one specific discipline and scholarship that is meant to be useful for academics working in other traditions. (If anything, the shrinking numbers of humanists and social scientists in American higher educational campuses provide a reason to write interdisciplinarily while also materially coagulating the different traditions into what should probably always have been shared interests.) Additionally, more and more scholarship today is also starting to reach out to a general audience, taking over the mantle from pop science books that endorse embarrassingly progressivist versions of technological advancement, often sanctioned and praised by Silicon Valley executives. Prominent examples of such critical work include Safiya Noble's Algorithms of Oppression (2018), Ruha Benjamin's Race after Technology (2019), and Cathy O'Neil's Weapons of Math Destruction (2016), all of which became (inter)nationally famous recently, for all the right reasons, and are cited heavily in this cluster.No doubt there is an immense need for such critique—let us call it technology critique—that takes contemporary technological narratives to task for their sociopolitical entanglements. Such technology critique, even when it is technically oriented, has a barely different job than conventional ideology critique; if anything, it is the mirror image of literary ideologiekritik. In historical literary criticism (think Raymond Williams, Max Horkheimer, Theodor Adorno, Étienne Balibar, and Fredric Jameson), ideology critique can be broadly understood as analyzing the relationship between sociohistorical ideological configurations and textual forms. Technology critique, then, precisely because its core is usually considered a material (e.g., machine) and not a solely symbolic (e.g., a text) artifact, offers its inversion. Now, scholars of science and technology know very well that, in Lacanian terms, machines can be symbolic systems, and that both the machine and the text are always material-semiotic at the same time; in fact, this was the road taken by literary scholars such as Kittler and Hayles when first staking a claim on studying new technologies, with their central arguments resting on the fact that the electronic computer is, in many ways at its core, an inscription machine. But regardless of whether they start from the material or the semiotic, critics always end up in both places at once. For example, Romi Ron Morrison's entry on "Flesh" in Uncertain Archives finds the materiality of flesh to also be an "unavoidable signifier and text" for the "history and reality of total expropriation of value from bodies and land" (249). Hayles's contribution on the "Unthought" in the same volume, on the other hand, finds the cognitive nonconscious—literally the grounds for what can be thought in the first place—to be utterly material in its neuroscientific, corporeal, and inscriptional-technological bases (546). Thus, these volumes under review exhort us to continue further delineating and complicating the relationships between different material-semiotic components of technology while keeping the core of this burgeoning interdisciplinarity in technology critique. In fact, technology, in its etymological sense, as the study (-logy) of technē (art, craft, practical skill) may already be ideological critique itself, and this call then is a mere return to technology in its originary sense; calling something an "institute of technology" may actually mean using the term study of technē, the closest to its original form. But the dominant mode of technology critique today still involves telling us what is wrong with contemporary technologies, and as hinted above, the causalities are often missing, or it is unclear what exactly is being critiqued. When exactly do we critique technologies themselves ("your AI is racist"), when are we blaming individuals (Musk, Bezos, Zuckerberg, or techbros in general, all rightly find fingers pointed at them in these volumes), and when exactly is a named system to blame (whether that is racism, patriarchy, settler colonialism, or racial capitalism)? While the cowardly (and correct) answer, of course, is usually all of the above, that cannot be a distraction from pointed questions that can still be asked in/of such scholarship today, questions that might eventually force us—the readers, the technologists, and the critics altogether—to carefully diagnose what we need to do for/in this glitching, if not already broken, world today.The role of AI as a technology might be an instructive case study here. After all, even within the work encountered in this piece, there is as much conversation about platformization, big data, and computation at large as about deep learning, neural networks, or advanced inference techniques normally associated with the term AI. Roughly, these books help us visualize the AI scene today; one can draw concentric circles, and AI will be the innermost: a tiny subset of soft- and hardware combinations that rely on a specific orientation of data sets and practices—which are often used to train AI—standing within a particular politicoeconomic, historical, and sociocultural framework of computational technologies. Put simply, conversations about technology today can often be conversations about computational technocultures, which in turn can often mean talking about big data, which itself is distinct from, but often enmeshed within, discourses of AI at large. These reviewed volumes make it clear that AI is not a coherent entity today; it shows up as either a technical assemblage at these different scales—of technologies, computational infrastructures, big data cultures, and so on—or a rhetorical maneuver toward vagueness as exercised by politicoeconomic power structures (Technoprecarious, 38–48; Your Computer Is on Fire, 51–70; Uncertain Archives, 65–75, 88–108). This "what is AI" question, when sincerely presented, shows AI to be a contingent, constructed category that relies on rhetorical maneuvers that birthed it in the first place. This was true in early days of AI in the mid-twentieth century, and it is still true now when the AI startup Kiwi Campus uses human labor to drive its remote-controlled Kiwibots around Berkeley, pretending to be "doing" AI, or as Sarah Roberts reminds us with her succinctly titled contribution in Your Computer Is on Fire, "Your AI Is a Human" (51). AI, it seems, at the end of the day is not just a technical problem; discursively, it signifies everything that can be possibly used to politically control and reinscribe global futures. And this possible reinscription is also a potentially productive moment to ask what is to be done. For example, when AI is scaled up, we get in the realm of big data, and the uncertainty invoked by the title Uncertain Archives suggests that big data is a wellspring of/for speculation, as all archives, per Jacques Derrida, have always already been (9, 11, 281, 403). In other words, these works teach us how technical assemblages are not set in stone; they retain in them a kernel of productive mischief.The fact that rhetoric, especially rhetoric that animates cultural conceptions and imaginaries, is a central component of what comprises "artificial intelligence"—or, for that matter, big data, or the tech sector, or computation at large—is worth noting here. It is this fact that fully crystallizes the role of humanities and, more specifically, literary criticism, in the context of this issue's concern with AI and literature. As the works reviewed here show, several material battles of technological futures are fought on the symbolic terrain. On the one hand, scholars of inscription are (or at least can claim to be) best suited to study the inscription machine that is a computer. On the other hand, we can study technical rhetoric through literary critical methods; in fact, the very choice to critique this instead of that is a priori a discursive-rhetorical imposition of material interests on us. As the AI expert Timnit Gebru (2022) lamented recently, [Our] research agenda has to be based on what tech bros decide to chase. They get $$$ chase X, X gets proliferated, then we spend all our time analyzing and mitigating the harms of X. We never wanted to build X, that was never in our research agenda, or interests, but now we have to spend all our time thinking about and dealing with X. . . . That leads to persecution, then you fight the persecution, then rinse & repeat, tech bros with $$$ have moved on to the next thing.This role of sociocultural factors in technological discourse thus essentially puts the onus of untangling naturalized notions about AI on scholars of language and inscription.Taking a lead from Noah Wardrip-Fruin—who in his brilliant contribution to Your Computer Is on Fire titled "You Can't Make Games about Much" shows how reskinning of games can be essentially a marketing maneuver that doesn't necessarily change the underlying gamic parameters, and therefore there are only a fini
Referência(s)