Automation and the Future of Work. Aaron Benanav
attainments and healthier lives, the labor share of income in G7 countries has fallen for decades (Figure 1.1).
Figure 1.1. Labor Share of Income, G7 Economies, 1980–2015
Source: OECD Compendium of Productivity Indicators, 2017, Chapter 1, Figure 1.8.
Such shifts signal a radical reduction in workers’ bargaining power. And the typical worker has faced harsher realities than even these statistics suggest, since wage growth has become increasingly skewed toward the highest earners: the infamous 1 percent. Growing gaps have not only widened between the average growth rates of labor productivity and of wages—which cumulatively causes the labor share of income to fall—but also between the growth rates of average wages and median wages—which evinces a shift in labor incomes from production and nonsupervisory workers toward managers and CEOs. The result is that many workers have seen a vanishingly thin slice of economic growth (Figure 1.2).27 Under these conditions, rising economic inequality will be contained only by the strength of redistributive programs. However, the “politics of social solidarity” have been weakening over time.28 Even critics of the automation discourse, such as economists David Autor and Robert J. Gordon, are disturbed by these trends: something has gone wrong with the economy, leading to a low demand for labor.29
Figure 1.2. Productivity-Wages Gap, OECD Countries, 1995–2013
Note: 1995=100. Employment weighted average of twenty-four countries, including Finland, Germany, Japan, Korea, United States, France, Italy, Sweden, Austria, Belgium, UK, Australia, Spain, Czechia, Denmark, Hungary, Poland, Netherlands, Norway, Canada, New Zealand, Ireland, Israel and Slovakia. For detailed information, see the OECD Economic Outlook. Source: OECD Economic Outlook, Volume 2018, Issue 2, Chapter 2, Figure 2.2.
Has runaway technological change been the cause of the low demand for labor, as proponents of the automation theory suggest? I will join critics of that theory in arguing that it has not. However, along the way, I will also criticize the critics—both for providing alternative explanations of a persistently low labor demand that apply only in high-income countries, and for failing to produce anything like a radical vision of social change adequate to the scale of the global labor underdemand problem, which has already beset the world economy for a long time and, due to COVID-19, is likely to worsen in years to come. It should be said from the outset that I am more sympathetic to the left wing of the automation discourse than to any of its critics.
Even if the explanation they offer turns out to be inadequate, the automation theorists have at least focused the world’s attention on the real problem of a consistently low demand for labor. They have also excelled in efforts to imagine solutions to this problem that are broadly emancipatory in character. The automation theorists are our late-capitalist utopians.30 In a world reeling from a global pandemic, rising inequality, recalcitrant neoliberalism, resurgent ethnonationalism, and the looming threat of climate change, automation theorists have tried to push through the catastrophe with a vision of an emancipated future, one in which humanity advances to the next stage in our history—whatever we might take that to mean—and technology helps to free us all to discover and follow our passions. That is true in spite of the fact that, as with many of the utopias of the past, these visions need to be freed from their authors’ technocratic fantasies as to how constructive social change might take place.
In responding to the automation discourse, the following chapters advance four counterarguments. First, I argue that the decline in the demand for labor of past decades was due not to an unprecedented leap in technological innovation, but to ongoing technical change in an environment of deepening economic stagnation. Second, I contend that this underdemand for labor has tended to manifest not as mass unemployment but rather as persistent underemployment. Third, I point out that the resulting world of poorly paid workers will continue to be accepted or even welcomed by elites, meaning technological advances will by no means automatically entail the adoption of technocratic solutions like universal basic income (meanwhile, even if UBI is introduced, it is much more likely that it will prop up a world of massive inequality than help dismantle it). Fourth, I explain how we might create a world of abundance even without the full or nearly full automation of production. I then project a path by which we might get there, through social struggle rather than administrative intervention.
Historically, major shifts in social policy have been adopted only under massive pressure, such as the threat of communism or of civilizational collapse. Today, policy reforms could emerge in response to pressure coming from a new mass social movement, aiming to change the basic makeup of the social order. Instead of fearing that movement, we need to see ourselves as part of it, helping to articulate its goals and paths forward. If that movement is defeated, maybe the best we will get is UBI, but that distributional reform should not be our aim. We should be reaching toward a post-scarcity world, a goal that advanced technologies will help us realize, even if the full automation of production is not achievable—or even desirable.
The return of the automation discourse has been a symptom of our era, as it was in times past: it has arisen when the gap between the supply and demand for jobs becomes so large, leaving so many individuals scrambling to find scraps of work, that people begin to question the viability of a market-regulated society. Even prior to the outbreak of COVID-19, the breakdown of the labor-market mechanism was more extreme than at any time in the past. This is because, over the past half century, a greater share of the world’s population than ever before came to depend on selling its labor (or the simple products of its labor) to survive in the context of weakening global economic growth rates. Our present reality is better described by near-future science fiction dystopias than by standard economic analysis; ours is a hot planet, with micro-drones flying over the heads of the street hawkers and rickshaw pullers, where the rich live in guarded, climate-controlled communities while the rest of us while away our time in dead-end jobs, playing video games on smartphones. We need to slip out of this timeline and into another.
A post-scarcity future—in which all individuals are guaranteed access to whatever they need to make a life, without exception—could become the basis on which humanity mounts a battle against climate change. It could also be the foundation on which we remake the world, creating the conditions in which, as James Boggs put it, “for the first time in human history, great masses of people will be free to explore and reflect, to question and to create, to learn and to teach, unhampered by the fear of where the next meal is coming from.”31 To find our way toward this post-scarcity future requires not only a break between work and income, as the automation theorists recognize, but also one between profit and income, as many do not.
Labor’s Global Deindustrialization
IF TECHNOLOGICALLY INDUCED JOB destruction is to have widespread social ramifications, it will have to eliminate employment in the service sector, which has absorbed 74 percent of workers in high-income countries and 52 percent worldwide.1 Purveyors of the automation discourse therefore focus on “new forms of service-sector automation” in retail, transportation, and food services, where “robotization” is said to be “gathering steam” with a growing army of machines that take orders, stock shelves, drive cars, and flip burgers. Many more service sector jobs, including some that require years of education and training, will supposedly be rendered obsolete in the coming years due to advances in artificial intelligence.2 Of course, these claims are mostly predictions about the effects that technologies will have on future patterns of employment. Such predictions can go wrong—as, for example, in the first week of January 2020, when three espresso-and-burger-slinging robotics