Salem Elzway is Assistant Professor of History at the Florida State University and a 2024-2025 Consortium Research Fellow.
My broad research program, “Race Against the Robots: Technopolitics and Inequality in Postwar America,” investigates how the design, development, deployment, and diffusion of automation shaped the political and socioeconomic contours of the United States after the Second World War. This history of automation is rooted in the reality that what technologies such as artificial intelligence (AI) and robotics can and cannot do, and what their a/effects and meanings are, is an all too human rather than mechanical affair. Rather than the determinism and teleology of the machine, it is political and socioeconomic institutions and structures that animate these actors to technologically engage their worlds in the ways they do. There is not a given or ‘natural’ reason why the uptake of artificially intelligent robots should create short, medium, or long-term hardships for workers whose jobs are technified or inequalities within and between communities affected by automation. “Race Against the Robots,” to play on the parlance of Science, Technology, and Society (STS), therefore, is designed to interrogate how artifacts, practices, and sociotechnical systems are not just socially constructed, but politically constructed. In this way, technologies may enable or amplify social phenomena such as economic growth and/or inequality, but on their own don’t cause them, as these are political problems emerging from political institutions, political structures, and political choices. In other words, a robot has never stolen a worker’s job!
The project emerging from this research, titled “Arms of the State: A History of the Industrial Robot in Postwar America,” integrates economic history, labor history, and state-centered political history with the scholarship of national security, political economy, and STS to provide the first archivally based scholarly history of the industrial robot. A variety of social actors from corporate managers, engineers, and government bureaucrats to intellectuals, labor leaders, and science fiction luminaries envisioned the industrial robot as the apex of automation that could be used as a geopolitical weapon to help win the Cold War. In the R&D facilities of the defense industry, the AI labs of MIT and Stanford, and the factories of Ford and General Motors (GM), technologists forged the robots of science fiction into the industrial robots of science fact. And industrial robots designed and developed by defense contractors as a direct replacement for humans in industrial settings did much more than manufacture material goods and disrupt labor markets. They functioned as a technology of worker discipline and as a totem of worker emancipation. They produced and reproduced racial difference inside and outside of both the factory and the laboratory. They played a crucial yet overlooked role in the early development of non-robotic AI. And they served as a dynamic symbol of/within the politics that shaped labor-management relations, the causation/correlation of technology and unemployment, and the evolution of the welfare state. These processes fueled a political environment that, with one hand, galvanized the power of the American state to actuate the development of AI and robotics, while with the other, attempted to contain what one economist and presidential adviser called the “human fallout” (i.e., social inequality producing effects) of automation. A significant intervention of the project is to enjoin the seemingly disparate subfields of labor history, political economy, and STS to map the sociotechnical networks of what automation was/is and who controlled/controls it so we can analytically navigate the technopolitics of the past, present, and possible futures of AI, robotics, and automation in practice.
Readers interested in the origins of AI and robotics; the cultural engine of science fiction; the scale and scope of the military-industrial-academic complex; the political economy of automation on the shopfloor; Silicon Valley entrepreneurialism; and the fall the American robot industry (qua the rise of its Japanese competitor) will find an abundance of materials and stories at the intersections of these domains. The project’s archive is composed of a wide range of primary and secondary sources, both published and unpublished including (but are not limited to): academic papers, congressional hearings, digital collections, federal government reports, newsletters, oral histories and interviews, periodicals, personal papers, private correspondence, trade journals, university archives and special collections, White House files, and works of science fiction. Notably, the book will be the first to use the George Devol Jr. Records and Joseph Engelberger Papers at the Benson Ford Research Center of the Henry Ford Museum in Dearborn, MI; records from the GM Heritage Center in Warren, MI; and the rarely-if-ever used Stanford Artificial Intelligence Laboratory Data Access and Retrieval Tapes (SAILDART), the digital repository of the lab’s communications and documentation from approximately 1972-1990. Expanding on the dissertation, the manuscript will include more comprehensive coverage of the early industrial robot industry during the 1950s-1970s; the development of the first US robot industry trade association in the early 1970s; the history of Carnegie Mellon University’s AI lab and Robotics Institute in the late 1970s and early 1980s; the critical relationship between American and Japanese AI and robotics interests in the 1980s; and the evolution of artificially intelligent robots and their uses after the 1980s.
Consortium-supported research for the project specifically, and my research agenda more broadly, took place at four sites: the Linda Hall Library; the University of Illinois-Urbana Champaign (UIUC) archives; the Smithsonian; and the American Philosophical Society (APS) collections. My research at the Linda Hall Library was geared towards obtaining rare materials related to the early history of the industrial robot industry including hard to find conference proceedings, industry-specific studies, and trade publications. One highlight of the documents reviewed is Mike “M. J.” Cooley’s paper “Industrial Robots: A Trade Union View of the Social and Industrial Implications” from the First Conference on Industrial Robot Technology held at the University of Nottingham, U.K. in March 1973. Cooley, an Irish-born engineer, trade unionist, and writer, directly challenged then prevalent optimism that “automation, computerization and the use of robotic devices” would automatically liberate people from “routine, back breaking, soul destroying tasks, and leave them free to engage in more creative worthwhile activities” (R20-223). Due to the constraints and pressures of a profit-oriented economy, and the two trends of increasing technological obsolescence and increasing capital intensity, Cooley argued that automation implemented in the name of accumulation would continue to degrade work and the worker alike while simultaneously ennobling critiques of technology as inherently brutal and inhuman. Rather than “fear technological change as such,” Cooley proclaimed, “We have a profound and historical responsibility to see that the products of our skill and ingenuity are used to provide the economic basis for a more full and dignified existence for the community as a whole, not merely to maximise profit for the few” (R20-228).
At UIUC, my research revolved around Heinz von Forester’s work on cybernetics at the university’s Biological Computer Laboratory (BCL), the history of the ILLIAC IV (the world’s first massively parallel ‘supercomputer’), and UIUC’s role in the military-industrial-academic complex. Von Forester and the BCL were particularly interested in cybernetic ‘autopoiesis,’ or the design of human-machine systems that could reproduce and maintain their own boundaries and functions from within (like organic cells). Researchers explored the theory through the construction of light-seeking and light-avoiding “cybernetic robots” akin to Grey Walter’s ‘tortoises’ (named Elmer and Elsie) and Norbert Wiener’s ‘moth’ (first named Palomilla and later called Felix), all of which demonstrated the possibility of artificial phototaxis. The lab also supported the publishing of a newsletter, The Cybernetician, that billed its content as “The Cybernetics of Cybernetics.” The most interesting materials I found on the ILLIAC IV related to the politics surrounding its funding source—ARPA and the Air Force—and anticipated uses—counterinsurgency modeling and nuclear weapons research. Like the March 4th movement at MIT in 1969 (and at other campuses across the country during the period), concern about the university’s involvement in the Vietnam War motivated faculty, student, and the broader Urbana-Champaign community to mobilize against ILLIAC specifically and war-related research generally. The proverbial pot boiled over when the university, to protect its relationship with military funders, founded a “Center for Advanced Computation” to house the ILLIAC, establishment of which bypassed traditional department channels, mandatorily classified work done on the machine (to the chagrin of Daniel Slotnick, director of the project and the computer’s designer), increased security clearance requirements, and allowed Department Defense control over the computer’s operating time. Protesters (and, to a degree, Slotnick) eventually won out as ARPA and the university agreed to move the computer to the NASA Ames Research Center in Silicon Valley.
My research at the Smithsonian served two purposes: first, to survey the institution’s collections related to machine tools and numerical control; and second for a parallel project on historian of technology David Noble’s curatorial work on an exhibit titled “Automation Madness” at the National Museum of American History that was slated for opening in Summer 1984. Materials related to the first, particularly those in the Robert G. Chamberlain Collection and the Max Holland Numerical Control Collection, explore the development and eventual economic contraction of the American machine tool industry. While the sources hew to the conventional story of ‘U.S. decline-Japanese ascent,’ documents from the National Machine Tool Builders Association (NMTBA) and The Diebold Group open a window on how U.S. corporate executives and military bureaucrats attempted to shape national technology policy, standardize automation, and address growing global competition from the late 1960s through the 1990s. Regarding Noble’s “Automation Madness” exhibit, archival finds range from the bureaucratically vapid to the personally vitriolic. Noble intended the exhibit to provide a unique and path-breaking narrative about the intertwined histories of labor, technology, and workers’ resistance to the machine of capital. The politics of the exhibit, both in content and practical execution, proved controversial, and Noble, whose career was characterized by principled political stances that got him into hot water at just about every institution he worked at, eventually alienated the Smithsonian bureaucrats who managed his curatorial purview. For more on Noble and this fascinating episode in the politics and public history of technology, I recommend my friend and colleague Richard Bachmann’s essay “In Search of an Exhibit that Never Came to Be: “Automation Madness” at the National Museum of American History, 1984.”
Last but not least, I visited the American Philosophical Society to begin scoping a future project on William Shockley and the politics of human, non-human, and artificial intelligence. Such a research domain is vast, and my time at APS mainly focused on how leaders of the National Academy of Sciences (NAS) navigated Shockley’s scientifically and politically untenable position on eugenics during the mid 1960s through early 1970s. Shockley repeatedly petitioned the Academy to sponsor formal investigations into the heritability and racial hierarchies of intelligence. When scientific bodies including the NAS declined to endorse his proscribed program, Shockley publicly attacked the scientific community, alleging that their ideological dogmatism amounted to a Western variant of Soviet Lysenkoism that stifled academic freedom and speech. Ultimately, an ad hoc committee of the NAS affirmed the general freedom of scientific research while concluding that a national research program on racial differences based on heritable intelligence was entirely unjustified and methodologically invalid. This chapter in the NAS’s history would be particularly relevant to scholars interested in how institutions and organizations navigate the politics of scientific controversy as well as scholars investigating the social construction of scientific credibility.
