Automated and algorithmic management is already here, invisibly shaping job quality for U.S. workers
Predictions about which jobs may be “lost” in the United States as employers implement new automation and algorithmic technologies often consider management tasks and roles as the least vulnerable to being entirely displaced by these new technologies. As Equitable Growth grantee and Research Advisory Board member David Autor explains, the tasks and roles that have been seen as less susceptible to automation so far are often those that can be described as nonroutine cognitive and social tasks, which are common in professional and managerial roles.
Of course, many forecasters do suggest that advanced automation and algorithmic technologies—aided and abetted by artificial intelligence—will one day advance to the point that they are able to replace managerial roles. But the loss of management roles to different kinds of automation and algorithms is not a far-off trend of the future. In many ways, it is already here, yet invisible—in part because of how U.S. companies use these new technologies to “offload” management tasks and roles to their lower-level employees, gig workers, and even customers.
U.S. companies use technology to automate, outsource, and offload workforce monitoring and management tasks in a variety of industries, which affects U.S. workers, as well as consumers, labor markets overall, and economic growth. Automation and algorithmic decision-making already affect U.S. workers throughout their employment relationships, starting with recruitment and hiring and extending throughout workplace environments.
Algorithmic management tools codify and extend management decisions in ways that likewise can affect many aspects of job quality, including wages, hours, and scheduling stability, as well as workers’ health and safety. What’s more, in the current U.S. landscape of weakened labor protections and rising corporate power, companies often use technological management tools in ways that worsen working conditions, reduce worker autonomy, and can have broader labor market effects connected to de-skilling, misclassification, and exercise of worker power.
Pervasive monitoring and automated management affect U.S. workers throughout the economy
Both platform-directed “gig” workers and workers classified as employees are impacted by automated and algorithmically driven management practices. Gig workers, who are frequently misclassified as “independent contractors” are particularly vulnerable to harms from these practices, often finding their entire work processes overseen by algorithms. But algorithmic management and monitoring practices also impact wage and salary workers in many industries, including retail, care work, call centers, and warehousing.
One common form this monitoring takes is employers’ use of algorithmically guided scheduling practices. Unpredictable or “just-in-time” scheduling is an example of an algorithmically guided practice that seeks to minimize labor costs and increase short-term efficiency, but which increases stress and economic precarity for workers.
Crucially, though, research also increasingly suggests such practices can be less efficient or productive overall. Research by Equitable Growth grantees Saravanan Kesavan, Susan J. Lambert, and Joan C. Williams, along with the University of Oregon’s Pradeep K. Pendem, examined the causal effects of scheduling practices in Gap Inc. retail stores, finding that moving to more consistent scheduling practices actually increases stores’ productivity and sales.
Algorithmic management in misclassified ‘gig’ work offloads invisible management work to U.S. workers and customers
Certain types of managerial decisions—such as those about a worker’s performance evaluation, pay, hours, and continued employment—are automated or algorithmically driven, while others are displaced to the workers themselves. In a recent working paper presented in the Academy of Management Proceedings, Princeton University sociologists Diana Enriquez and Janet Vertesi analyzed interviews of gig workers who worked for Uber Technologies Inc., Amazon.com Inc., and Lyft Inc. The researchers describe a process of “automation by omission,” in which the companies’ platforms are designed to automate key parts of what would otherwise be a “middle management” layer of oversight, but in a way that functionally (and invisibly) offloads other middle management tasks to the workers themselves.
These middle management tasks include constant decisions about how to optimize earnings, purchase and manage equipment and other costs, and even crowdsourcing ways to improve performance with other drivers in online forums. These companies also use technology to shift management functions directly to customers, whose feedback or ratings of workers may directly impact their performance evaluation with no manager or mediating actor.
As a result, Enriquez and Vertesi find that the workers “bear the costs associated with representing a platform and themselves when there is any contrast between the projected expectations of the customer and the completed service.” Platform gig services give customers “an upper hand in transactions” relative to the workers, as a 2021 analysis of algorithmic management in work explained, and company-side management tools are predominately limited to aspects such as the integrity of payment transactions and monitoring workers’ performance.
An example of this is documented in the recent report, “At the Digital Doorstep: How Customers Use Doorbell Cameras to Manage Delivery Workers,” by Aiha Nguyen and Eve Zelickson at the independent research organization Data & Society. Their report shows how customers display three broad types of management or supervisory behavior over delivery drivers: monitoring, instructing, and punishing. By classifying workers as independent contractors and giving customers direct managerial input on them, in addition to the company-side monitoring and control, companies are thus able to enjoy the benefits of a highly monitored workforce without the direct managerial costs—or regulatory and legal accountability—that would otherwise accompany such detailed management.
Further research and new policy actions can protect U.S. workers from visible and invisible algorithmic management decisions
One challenge in understanding the full scope of present and potential harms from these practices is that U.S. employers often use technology in managing workers in ways that are difficult for external researchers, policymakers, and regulators to track, and are generally not required to share that information publicly or with regulators.
Increasingly, however, academic researchers, policy experts, and worker groups are working to expand our knowledge of how new forms of technology affect workers and the economy to bring this “invisible work” into research and policy conversations. Researchers have already outlined many of the ways that algorithmic management and other data-driven technologies harm workers and shape working conditions, as well as key areas of interest for policymaking.
But more work remains to be done. As described in our 2023 Request For Proposals, the Washington Center for Equitable Growth is interested in research that seeks to answer questions around how employers use automation and other new technologies in the workplace and how these decisions interact with and are affected by the broader regulatory and policy landscape.
This research will be assisted by new policy efforts to bring accountability to these practices. Just this month, new policies went into effect in California giving workers employed by large businesses more information about how their employers are surveilling them and using data about them—an important first step of many in bringing labor protections into the 21st century.
Cities and states also are addressing specific practices common within algorithmic management, such as just-in-time scheduling practices. And at the national level, regulators and policymakers are likewise examining how to protect workers from the potential harms of automated decision-making and increase accountability and oversight in U.S. workplaces. Taken together, these and other promising policy avenues can help mitigate these harms to U.S. workers, alongside a growing body of research examining those dangers and the efficacy of those technologies in U.S. workplaces.
