Thinking with Robots
2023; Duke University Press; Volume: 95; Issue: 2 Linguagem: Inglês
10.1215/00029831-10575162
ISSN1527-2117
Autores Tópico(s)Neuroethics, Human Enhancement, Biomedical Innovations
ResumoThe word robot first appeared in English in 1922, in a translated New York production of the Czech writer Karel Čapek’s 1920 play, R.U.R. (Rossum’s Universal Robots). There, it echoes the words for labor (robota) and laborer or serf (robotnik), from a range of Slavic languages, including Čapek’s Czech, in which robota specifically suggests forced labor or drudgery. In Čapek’s play, an international success translated into over thirty languages within three years of its premiere, robots are artificial people created by “philosopher” and “scholar” old Rossum from a process involving “protoplasm,” a synthetic organic material (Čapek [1920] 2004: 5). Although old Rossum did not originally intend robots only to work, his son capitalized on his father’s creation by “chuck[ing] everything not directly related to work” in the robots’ psyches and mass-producing them to serve as replacements for human laborers and soldiers (9). The play opens in Rossum’s factory, located on a far-flung island, at the height of the company’s success; by the end of the play, ten years later, robots have revolted and, in an attempt to become more like the humans they were intended to serve, they have seized power by killing all humans throughout the world except one. While the complex formula for robot reproduction was destroyed by humans at the factory before their deaths, the play nevertheless ends with a Christian fantasy of reproduction and dominion. Primus and Helena, a male and female robot couple now endowed with human “souls” and thus, the play implies, the potential to create life, leave the factory to repopulate the world. Alquist, the sole human survivor, calls them “Adam” and “Eve” as he pushes them out the door of the factory, shouting, “Life shall not perish!” (84).It’s hard to overstate both how strange Čapek’s play is and, in light of one hundred subsequent years of sci-fi tales about robot revolts, how clichéd it seems to a contemporary reader. This uncanny familiarity is due in no small part to its profound influence not only on twentieth- and twenty-first century science fiction but also on the fields of robotics and artificial intelligence (AI). While each of the works reviewed here takes a distinct approach to this influence and legacy, they all position R.U.R., and thus literature, as central to the robot canon. Stories about how robots are soon going to replace “us” and take “our” jobs have long been a fixture in the United States, but these books ask readers to consider exactly who is included in such proclamations and precisely what fuels this persistent anxiety. They join other recent works of scholarship such as Minzoo Kang’s Sublime Dreams of Living Machines (2011), Gregory Jerome Hampton’s Imagining Slaves and Robots in Literature, Film, and Popular Culture (2015), and Scott Selisker’s Human Programming (2016) in considering the role of automation and the figure of the robot in producing conceptions of the human, a dynamic at the foundation of Čapek’s play that I discuss in the first half of this review as central to each of these books. Also important to Čapek’s play, and thus to the prevailing ideas about robots these books examine, is a fixation on the bodies of robots. In the second half of this review, I push at the limits of this fixation by thinking through what these books might offer to our understanding of robots that aren’t exactly embodied in this way: bots on social media and cyborgs.Despina Kakoudaki’s Anatomy of a Robot: Literature, Cinema, and the Cultural Work of Artificial People discusses the performance history and reception of R.U.R. in some detail, tracking how the robots’ costumes became more metallic and artificial looking throughout the first decade of its production. Despite its title, Kakoudaki’s transhistorical study is not limited to discussions of robots per se. Instead, it emphasizes that, while we can trace the first uses of the word robot to R.U.R. (Čapek’s brother, the artist Josef Čapek, coined the term and suggested Čapek use it in his play), narratives of artificial people have a much longer history in Western cultures. Anatomy of a Robot tackles the big question of why such narratives continue to fascinate despite this long history and the repetitive nature of these tales. Kakoudaki organizes her book thematically around four “core feature[s] . . . of the discourse of the artificial person”: artificial births, mechanical bodies, enslavement, and philosophical questions surrounding the idea of artificiality itself (26). Drawing on a large variety of examples, Kakoudaki deftly places contemporary works of literature, television, and film in conversation with older narratives and myths about artificial people. The first chapter, focusing on artificial births, draws connections between Mary Shelley’s Frankenstein (1818) and creation myths such as the story of Adam and Eve and the birth of Pandora, ancient and modern stories about the golem, and early science fiction films such as Fritz Lang’s Metropolis (1927) and Boris Karloff’s Frankenstein (1931). Other chapters discuss modern and contemporary robot stories alongside Renaissance art and anatomical drawings, early experiments with electricity, and eighteenth-century narratives of imperialism and the slave trade, such as Jonathan Swift’s Gulliver’s Travels (1726) and Olaudah Equiano’s The Interesting Narrative of the Life of Olaudah Equiano, or Gustavus Vassa, The African (1789). This expansive organizational scheme is held together by Kakoudaki’s illuminating close readings, which show how similar ideas link stories of artificial people across time, genre, and medium. Crucially, these close readings return to the precarious legibility of the category of the human, which Kakoudaki shows has so often depended on an opposition to those deemed in- or nonhuman.But not all of Kakoudaki’s claims are transhistorical. While the first two chapters emphasize larger philosophical questions about birth and death, growth and decay, control and freedom, the focus on embodiment suggested by the “anatomy” of her book’s title leads her in chapter 3 to historically specific arguments about race and labor. This chapter emphasizes the connections of robot stories to the modern history of the transatlantic slave trade. Kakoudaki discusses how robot stories, including Čapek’s R.U.R., as well as Isaac Asimov’s robot stories of the 1940s and his novella The Bicentennial Man (1976), operate as “allegories of otherness” that, in twentieth- and twenty-first-century contexts, revolve around racial difference and the legacies of chattel slavery (117). The chapter details an effect Kakoudaki terms metalface, or how “the metal exterior of the robot functions as a site for projecting numerous kinds of difference,” a fantasy of “racist epistemologies” in which “one may be able to tell where a person fits in a social hierarchy just by looking at them” (117). As in Čapek’s play, these modern and contemporary texts almost invariably include scenes of robot revolt in which robots develop an awareness of their enslavement and rise up against their enslavers. Kakoudaki reads such scenes as liberal fantasies of self-consciousness that occlude the actual political and social mechanisms that enforce slavery. The liberalism at the root of these stories of robot revolt relies on a stable conception of the human, one that inevitably demands a nonhuman other against which to define itself. But Kakoudaki’s fourth and final chapter suggests other, more potentially optimistic possibilities. This chapter focuses largely on modern and contemporary texts such as Ridley Scott’s Blade Runner (1982), the twenty-first-century version of the television show Battlestar Galactica (2004–9), and Kazuo Ishiguro’s Never Let Me Go (2005) that destabilize the relationships between interior and exterior, artificial and real, and object and subject suggested by the texts in the previous chapter. Kakoudaki argues that these texts may point to “new understandings of how humanity might be defined” that do not rely on an opposition to the nonhuman for their legibility and that question divisions between “artificial” and “real” people altogether (174).Kakoudaki discusses Čapek’s R.U.R. in the middle of her study, positioning it as an inflection point in stories of artificial people that would pave the way for the many robot stories to follow. In contrast, in The Sound of Culture: Diaspora and Black Technopoetics, Louis Chude-Sokei includes an extended discussion of the play in the book’s first chapter to initiate an argument about the implied opposition between Black people and technology at the foundation of British and American literary modernism. Chude-Sokei’s study focuses broadly on the relationship between race and technology, ranging from nineteenth- and twentieth-century science fiction such as Čapek’s play, Samuel Butler’s Erewhon; or, Over the Range (1872), and William Gibson’s Neuromancer (1984) to Donna Haraway’s well-known essay “A Cyborg Manifesto”; the histories of jazz, reggae, and dub music; and the work of Caribbean studies scholars such as Éduoard Glissant, Wilson Harris, and Sylvia Wynter. Like Anatomy of a Robot, The Sound of Culture investigates discourses about the legibility of the human, but Chude-Sokei is less interested in artificial people or robots per se. Instead, he documents how “technology has always been racialized or articulated in relationship to race” (2).This wider remit allows for incisive arguments about, for example, the legacies of such terms as master and slave in engineering and robotics. It also makes space for several brilliant pairings, such as comparing R.U.R.’s robots to robotic characters in Jean Toomer’s roughly contemporaneous work “Rhobert,” a story collected in Cane (1923), and his play Man’s Home Companion (1933). Chude-Sokei also includes a reading of Čapek’s satirical 1936 science fiction novel War with the Newts, which offers a more explicit allegory of racial violence than R.U.R. and mocks fascism and colonialism more openly. He focuses throughout on what he terms black technopoetics, which loosely describes “the self-conscious interactions of black thinkers, writers, and sound producers with technology” and unites his many objects of study (11). While music history is a thread that weaves in and out of chapters, The Sound of Culture is less a work of musicology or sound studies than an indication of what is possible when music, sometimes a neglected medium in cultural studies, is considered as a kind of technology. For Chude-Sokei, music is important to his arguments because “music has been the primary zone where Blacks have directly functioned as innovators in technology’s usage” (5).The first two chapters of The Sound of Culture contextualize a wide variety of American and British nineteenth- and early twentieth-century science fiction in relation to the history of the phonograph, minstrelsy and Uncle Remus songs and tales, Italian futurism, and Norbert Wiener’s work in cybernetics. The historical depth of these chapters convincingly suggests a new genealogy of Anglophone science fiction, one in which “race, racialization, slavery, colonialism, and technology become a part of the DNA of science fiction from its genesis” (79). The final two chapters turn to more contemporary works, focusing mostly on late twentieth-century cyberpunk, posthumanism, and theories of creolization rooted in a Caribbean intellectual tradition, or what Chude-Sokei refers to as a “Caribbean pre-posthumanism” (179). These chapters, connected by their interest in the concept of hybridity, are more abstract and theoretical than the previous two. Nevertheless, Chude-Sokei makes unexpected and productive connections, asking readers to reconsider the well-trodden ground of cyberpunk by focusing on the genre’s interest in reggae and dub music as more than “mere exoticism,” and unfolding a prehistory of Sylvia Wynter’s work rooted in the Harlem Renaissance, Black surrealism, and Afrofuturism (130). Like Kakoudaki’s final chapter, the second half of Chude-Sokei’s book attempts to imagine the human and the technological without the binaries that have historically—and with world-historically destructive consequences—structured these categories. Taking his cue from Wynter, Chude-Sokei develops in these chapters a “counterhumanism” that insists on the fusion of these categories, on the organicity of the machine and the artificiality of the human (224).Returning to the robota at the robot’s origin, Jennifer Rhee’s The Robotic Imaginary: The Human and the Price of Dehumanized Labor shows how, in film, art, and literature, as in the AI and robotics industries, robot labor is consistently imagined as devalued labor. In the book’s introduction, Rhee positions Čapek’s R.U.R. as foundational to this robotic imaginary, showing how “dehumanization largely occurs at the site of labor” in the play (17). This insight structures Rhee’s book, which is organized around the different kinds of devalued labor that robots perform: care labor, domestic labor, emotional labor, and drone labor, the military work of extending and maintaining the authority of the national security state through drone strikes. Rhee’s feminist critique highlights how these forms of labor are gendered and associated with or historically largely performed by women. Rhee also shows how the robotic imaginary she tracks assumes that the category of the human is recognizable, knowable, and somehow universal. Like Kakoudaki, Rhee is skeptical of these dynamics, and she shows how, in taking the category of the human as a given, the “robotic imaginary” excludes, silences, and devalues nonwhite and nonmale subject positions. The book turns instead, like Chude-Sokei, to Glissant’s concept of the opacity of the human, arguing for an understanding of the human that does not depend on recognizability or on such “extensive erasures of human experience” (6).The first three chapters of The Robotic Imaginary make the case that the devaluation of the labor of social reproduction relies on the dehumanization of those who typically perform this labor, including the gendered robots Rhee discusses. These chapters include illuminating readings of the film Her (2013; dir. Spike Jonze), the 1972 novel The Stepford Wives (Ira Levins) and its 1975 film adaptation (dir. Bryan Forbes), and Philip K. Dick’s novels Do Androids Dream of Electric Sheep? (1968) and We Can Build You (1972), as well as contemporary and historical examples from the robotics and AI industries, such as the AI therapist ELIZA and Kismet, a robot designed to look like a cute creature and express a range of human-like facial expressions. Rhee ends each chapter with a discussion of contemporary robot art; the pieces she selects exemplify the potential of the robotic imaginary to reconfigure ideas about the human, recontextualizing each chapter’s discussions of devalued labor and the boundaries of the human.The book’s fourth and final chapter pivots to a compelling discussion of drone labor that, Rhee argues, reveals just how much “militarization and reproductive labor are intimately entangled” (133). Drone warfare is emblematic for Rhee of the racial dehumanization at work in the robotic imaginary overall: it reveals “just which humans have been imagined to be the beneficiaries of . . . reproductive labor, and just which humans have been imagined to be disposable” (134). This chapter discusses works of drone art, such as Teju Cole’s Seven Short Stories about Drones (composed for Twitter) and the art installation #NotaBugSplat by a group of Pakistani and US artists, as well as the use of drones by the US military abroad and by US police forces at home. Rhee argues that these works of drone art both reflect and, through their insistence on a conception of the human “constituted not through the known, but through the unknown and the unfamiliar,” challenge the labor of racial dehumanization at the foundation of drone warfare (173). This fourth chapter thus adds a challenging perspective to her account of contemporary robot imaginaries: it may be, after all, in the increasingly technologically intertwined domestic and military spheres that robotics most powerfully shape the everyday worlds of US empire building. As Rhee shows, the labor of social reproduction and the labor of militarized killing are not just entangled—they are two sides of the same coin.The most recent of the books under review here, Neda Atanasoski and Kalindi Vora’s Surrogate Humanity: Race, Robots, and the Politics of Technological Futures, makes this connection an explicit part of its argument. Like the other titles discussed in this review, Atanasoski and Vora’s work points out the reliance of robotics discourse on classically liberal definitions of the human that depend on the exclusion of the nonhuman other. Focusing more explicitly on our present moment and building on Lisa Lowe’s work on liberalism and racial capitalism, they term this dynamic technoliberalism, “the political alibi of present-day racial capitalism” (4). Technoliberalism, they emphasize, defines the freedom of the liberal subject against the unfreedom of “degraded and devalued others” (4). Here, the authors’ language of surrogacy signals the feminist stakes of the anxieties about labor and substitution that have been discussed in all four books under review here. Unpacking the persistent fantasy and fear of human obsolescence, undergirded by long histories of racial violence, they argue that technology today exists in a surrogate relation to the human. They show how discourses of technoliberalism as articulated by the popular news media, scientists, politicians, the military, and industrialists and tech company CEOs position robotics and AI as substitutes for human labor, especially for forms of degraded labor. This process of substitution, in turn, defines what liberal subjecthood is: freedom from such labor. As they write, “The liberal subject is an effect of the surrogate relation,” a subject position made available (for some) through this dynamic (5). In this way, following from the work of Hortense Spillers, they show how this surrogate human effect is “the racial ‘grammar’ of technoliberalism” (5).Atanasoski and Vora’s book is located within feminist science and technology studies; unlike the other authors reviewed here, they are not professors of English or literature, and they are less concerned with analyzing literary or filmic depictions of robots than they are with accessing the larger discourses that animate contemporary “robotic imaginaries,” to use Rhee’s phrase. Nonetheless, Čapek’s R.U.R. also finds a place in their argument. After an introductory discussion of desires for and anxieties about technological enchantment, Atanasoski and Vora situate Čapek’s play within a longer history of technoliberalism in the twentieth- and twenty-first-century United States. Here, alongside discussions of Cold War automation and deindustrialization and Donald Trump’s promise to build a border war to prevent immigrant labor, the authors show how the color-blind technoliberal fantasies about freeing humans from drudgery depicted in Čapek’s play are deeply related to their supposed opposite, fascist ethnonationalism centered on protecting white labor.Chapters 2–4 take on contemporary manifestations of technoliberalism: the sharing economy, human labor specifically disguised as machine labor, and the development of social robots in the field of human-robot interactions. These chapters include incisive analyses of a wide-ranging set of examples, including “collaborative” robots that require human supervision, such as Baxter (built for factory work) and Botlr (built for the hospitality industry); feminist collaborative GynePunk’s 3D-printable speculum; gig economy services such as AlfredClub, through which users can arrange for an “Alfred” or personal butler to perform household tasks without ever interacting with this person directly; and Kismet, as mentioned above, as well as the social robot Omo, a machine designed by artist and engineer Kelly Dobson as a critique of companion robots. Chapters 5 and 6 turn to the automation of warfare, focusing on drones and killer robots, including some of the examples of drone art that Rhee also discusses, such as #NotaBugSplat. Atanasoski and Vora argue that these military technologies construct killable populations of people as “targets,” providing “a technoliberal postracial update to present-day imperial conquest” (138). They also focus on how these forms of racial violence work through and against their supposed limits. For example, they emphasize that drone operators who care about their targets are construed as having failed at their jobs, and that discussions about banning killer robots often revolve around the idea of cutting humans, and thus the supposed possibility of empathy, out of the loop. The book ends with a fascinating epilogue about sex robots and ideas about “feminist AI” that seek to counter the simulation of consent built into sex robotics. In demonstrating how such seemingly feminist approaches end up “expand[ing] the category of intelligence without necessarily disrupting its value,” Atanasoski and Vora point the way toward challenging the seeming self-possession gained through the surrogate effect (196).Each of the “robot books” reviewed here emerges from a distinct disciplinary and methodological perspective. Nonetheless, each returns to the tenuous boundary between the robot and the human, interrogating the powerful dynamic by which the category of the human is persistently reasserted against the nonhuman other that the robot represents. These classifications are often negotiated through bodies and embodiment. Many of the fictional and actual robots discussed in these books, as in Čapek’s R.U.R., are human-like androids, defined by the humanness of their form and the inhuman matter that composes it. These books’ interest in the bodies of robots unfolds in concert with the attention they pay to the legacies of racial and gendered violence that have long energized the figure of the robot. Among these titles, Kakoudaki’s Anatomy of a Robot focuses most explicitly on embodiment as such and, along with the first two chapters of Chude-Sokei’s The Sound of Culture, most closely attends to how these kinds of robot bodies are imagined and experienced. But nonandroid robots such as Kismet or drones, both of which appear in Rhee’s The Robotic Imaginary and Atanasoski and Vora’s Surrogate Humanity, of course, are also embodied. This is not only because they have touchable metal “bodies” or because they are mobile and animate and so project an illusion of autonomy but also because, as these works show, robots often function as figures or stand-ins for ideas about human bodies. As Atanasoski and Vora argue most explicitly—but as all of the works reviewed here discuss in one way or another—even those robots that don’t appear to have any physical resemblance to humans nonetheless threaten (or promise) to substitute for them. Across these titles, forms of robot embodiment are determined by these two entangled dynamics: one of resemblance between robot and human bodies, with the crisis of legibility such resemblance invites, and one of substitution, or the power that robot bodies, whatever they look like, have to replace human labor.But what about those robots that don’t have bodies in this way? In what remains of this review, I offer two examples of the robotic that exist in tangent to the many examples discussed in these robot books. It is not exactly correct to say that the examples of the robotic I discuss below are “embodied,” but that is not exactly incorrect, either. The first example I discuss, bots on social media, illustrates the intellectual reach of the robot books discussed here. Though none of these books addresses this kind of robot at any length, the example suggests how their arguments and insights might help us understand aspects of contemporary culture, such as social media, that we may not have considered within this context before. The second example, the reclamation of the retrofuturist figure of the cyborg by disabled writers and activists, helps us better understand the limits of the robotic imaginaries these books reveal. These writers and activists are responding to some of the same dissatisfactions with liberalism that we have seen in the four books under discussion here; however, their resistance to liberal humanism turns to the figure of the robot in a different mode, and to a different end. While discussions of robot anatomy, race and technology, dehumanized labor, and the surrogate effect turns on distinctions between the human and the robot, those who claim a cyborg identity refuse such distinctions. This refusal directs our attention from ideas about embodiment to lived experience, from robots as metaphors or allegories or substitutes for the human to everyday life as a cyborg.I was at first surprised to see only passing discussion of bots in these four books, especially given growing anxiety in recent years about the geopolitical influence of bots on social media. But this omission is less surprising when we consider that, as software programs that perform automated tasks, bots have an ambiguous relationship to the forms of robot embodiment central to these books. The term bot in this context comes out of the protosocial networks of the precommercial internet, specifically the Usenet group alt.mud. This group was an early internet forum for discussing text-based multiplayer real-time virtual worlds (originally multiuser dimensions or dungeons), or MUDs. Popular in the 1980s and 1990s, MUDs combined elements of role-playing games, interactive fiction, and online chat. Players learned about the virtual world by reading descriptions of rooms, other players, and objects, and they interacted with one another and their surroundings by typing commands. The goals of these games often involved completing quests and roleplaying. By the late 1980s, many MUDs were populated with a variety of automated players—or, as they were referred to at the time, robots or clients—designed to mimic human players, provide information about the virtual world, or complete game-related tasks for their human creators. On January 23, 1990, reacting to the increasing number of bots in the MUDverse, user Heresiarch sent a message titled “bot-haters unite!” to the alt.mud Usenet forum. Shortening robot to bot in what the OED records as the first use of the term to mean an automated computer program that attempts to mimic human activity on a social network, Heresiarch (1990) began the message, “The following consists of a general flame against bots.”1 Heresiarch states that, while they didn’t mind one of the original MUD bots, Gloria, because “she” “was cute and entertaining and gave out useful information,” “now tinymud has a dozen gloria clones with no personality running around. . . . i’m sick of this.” They end the message with a declaration of “BOTWARS!”: “all bot-hating mud citizens are hereby encouraged to do everything possible to fuck with the tiny little . . . monsters.”Heresiarch’s description of their hatred for bots recapitulates the terms of racialized and (here, explicitly) gendered violence that these robot books have powerfully connected to the figure of the robot. Indeed, many of the bots that populated the MUDverse in the late 1980s and 1990s were explicitly gendered female. For example, another well-known bot, Julia, described itself as “5′1″ tall, weigh[ing] 123 lbs, with close-cropped frizzy blond hair and dark brown eyes,” as a “gossip,” and as a “secretary at a University”; Julia also tended to reference “her” “PMS” and “period” about two days a month (Foner 1993: 11, 7, 26, 16). Julia performed many useful functions in these games, especially for new players, such as offering assistance with directions and help commands, conveying messages to absent players, and recording conversations between players.2 As Rhee writes in The Robotic Imaginary in relation to ELIZA, the famous therapist chatbot and a precursor to Gloria and Julia, “Female AIs . . . frequently provide services associated with feminized care labor positions” (36). Moreover, such labor and those who perform it are often devalued and degraded. The Julia bot, for example, was the subject of a “kill Julia contest” in 1990 in which players competed with one another to concoct the most “creative” ways to kill Julia (killing players in a MUD was possible but not generally encouraged). Several of these kills, unsurprisingly, involved imaginations of sexual harassment or assault (Dirque 1990).Early bots such as Gloria and Julia are direct precursors of AI “assistants” such as Apple’s Siri, technology briefly discussed in Rhee’s and Atanasoski and Vora’s books. However, their legacy also lives on in what may be the contemporary world’s largest robot populations: the bots that operate fake accounts and profiles on social media platforms. These “social bots” are designed to spread (mis)information online, and they can operate either autonomously or semiautonomously with some human oversight. Estimates of the proportion of social media accounts that are actually bots range widely, as it’s often difficult to tell (especially using automated methods) which accounts are bots and which are people.3 It’s also often difficult to tell how influential social bots really are online; while bot accounts on Twitter, for example, can tweet and retweet with much more frequency than human accounts can, if those bots do not interact with very many people or if their tweets do not have much reach, the sheer number of times they tweet is not necessarily meaningful. These limitations aside, from frequent discussions in government and in mainstream med
Referência(s)