The MIT Task Force on the Work of the Future has been charged with examining the implications of technology-related transformations for employment. How will transformations associated with AI and automation affect the future of work? Are people right to fear that a growing army of robots is coming to take their jobs? What repercussions will the changing nature of work have, and for whom? In a tumultuous year when crises relating to governance, racism, misinformation, and a global pandemic have joined the more slow-burning crisis of climate change to batter the United States, some might ask whether questions about the work of the future are among the most pressing for us to address. The answer is yes. The other major crisis of our time—expanding economic inequality—underwrites and magnifies the fault lines of all other crises. The impacts are cumulative, and attention to the changing nature of work can point out the kinds of additional pressures that may lie ahead.
The disproportionate death rate from COVID-19 among African Americans is one example of this intersection of today’s growing inequalities. The effects of racism have become bound up with disproportionate exposure rates for groups with high numbers of “essential workers.” What it means to be an essential worker, however, has shifted with the changing nature of jobs in what David Weil, Dean and Professor of the Heller School for Social Policy and Management at Brandeis University, has called the “fissured workplace,” settings where employees find themselves increasingly outsourced, sub-contracted, working part-time or on demand, and with less leverage and fewer worker protections (Weil, 2014; Greenhouse, 2019). It is linked to the reality that loss of employment in the United States often means loss of health insurance as well as other essential benefits, while the uneven fallout of the 2008 financial crisis has left many with fewer resources and government supports to weather such crises. The result has been increasingly extreme patterns of compounded inequalities, affecting a disproportionate number of people of color.
I come to questions about the future of work as a cultural anthropologist. Among other things, anthropologists explore how people’s interpretations of the world are key to analyzing broader economic and social patterns.1 Consequently, as an anthropologist thinking about robotics and artificial intelligence (AI), I begin with the question: What do the robots symbolize? Over the last few years, worries about employment and displacement by robots and AI have become pervasive in popular discussions in the United States. While robots might be associated with many things (ranging from futuristic technophilia to dystopianism in Hollywood movies, for example), robots in the United States increasingly appear as symbolic encapsulations of broader anxieties about the changing nature of contemporary work. The observation that public discussion in the United States has gone from largely ignoring questions of technological transformation a decade ago to overexaggerating its effects by imagining robots as capable of far more than is currently true confirms that more is at stake than simply “robots.”
As a result, attempts to rein in public anxiety by debunking exaggerated claims about technology (either by challenging technologically determinist arguments or by noting that AI cannot do everything that boosters claim or critics fear) risk missing the point. Anxieties about robots are potent because they symbolically condense two key realities. First, middle- and working-class Americans do have ample cause for worry due to the expansion of increasingly precarious forms of employment which began long before COVID-19 generated widespread job loss. And, second, the reasons for these broader economic transformations are often opaque, making it tempting to focus on their material manifestations, in other words, the technologies themselves.
Tellingly, anxiety about robots is not ubiquitous across all countries. In Scandinavian countries, for example, robots are more commonly seen as welcome adjuncts to labor (Goodman, 2017). The reasons why are both obvious and telling. The United States’ thin and embattled social safety net makes the consequences of job loss and insecurity far more dire here than in those wealthy countries where citizens can rely upon nationalized health insurance, guaranteed sick leave, subsidized child and elder care, and strong supports for education and retraining.
The goal of this brief is to move beyond “robot as Rorschach test” for job anxiety and explore some of the systemic reasons for changes in employment and the potential impacts of job displacement, particularly on more vulnerable individuals, families, and communities in the United States. History and ethnography3 can help us with both. My own research focuses on working-class and lower middle-class populations, or those whom academics view as most susceptible to job displacement from robotics and AI. It builds upon research on everyday experiences of job loss historically among multiracial working-class populations in the Calumet region of Southeast Chicago and Northwest Indiana, once one of the largest industrial corridors in the world (Bensman and Lynch, 1987; Walley, 2013; Boebel and Walley, 2017; Walley et al., 2020). In particular, I am concerned here with lessons to be learned from past experiences of technological and work transformation in formerly industrial communities. These studies suggest that understanding work transformation entails addressing not only the numbers and types of jobs lost or gained, but also the forms of sociality, identity, and meaning associated with that work. Such studies also indicate how contemporary work displacement linked to robotics and AI is likely to exacerbate prior rounds of job displacement that have already had pronounced impacts.
I argue that the lessons to be learned from this historical and ethnographic perspective for policymakers are twofold. First, there is a strong need for public policies that discourage companies from enhancing their profitability by using technology to displace workers or downgrade skills and that instead encourage the use of technology to increase productivity in worker-supportive ways. Second, this scholarly work points to the need for a dramatic re-envisioning of social safety nets, an area where comparative research across countries can offer valuable insight. In sum, looking to the past and thinking about the ways that technology is embedded within and emerges from social relationships, rather than thinking in more technologically narrow ways, can help hone the questions we need to ask to plan for collective futures.
This discussion, however, first requires examining the disparate assumptions built into definitions of “social class,” “skill,” and “knowledge,” in order to counter the confusion and misconceptions that can result. During the 20th century, social class was generally determined on the basis of such factors as occupation, income, positioning within systems of economic production, and (sometimes) more social intangibles. However, this perennially fuzzy concept has become even more indeterminate. Many journalists and academics now define “social class” on the basis of whether or not individuals hold four-year college degrees. Although easier to measure, this redefinition of class transforms it into a binary metric—either you have a four-year college degree or you don’t. Those with a college education are now often assumed to be “elites” and those without one to be working class—a sleight of hand that conceptually erases the “middle class” once seen as quintessential to the American social landscape. This reconfiguration creates such anomalies as lumping together hedge fund billionaires with kindergarten school teachers among the “educated” and suggests that a college dropout like Bill Gates has something in common with a factory worker without a degree. It also generates an (overly) expansive umbrella for who might be considered “working class” (60% of non-Hispanic whites, for example, lack a bachelor’s degree) (U.S. Census Bureau, 2020; Metzgar, 2016; Walley, 2017). When pundits use the term “working class,” for example, it is often unclear who is meant—the industrial or former industrial workers central to its 19th-century definition? Rural farmers? Suburban business owners with some college? Office workers or managers? The impoverished?
Similarly, “skilled” work is now commonly assumed to be that which requires formal higher education, erasing long histories of skilled manual labor or other forms of craft or knowledge. The slippage in definitions of class and skill suggests both changing social realities on the ground and the need to examine the (potentially misleading) assumptions we bring to these discussions.