The Power of Witnessing Work: AI and What Humans Do for Each Other

• In Event: 4410 - Artificial Feelings: The Politics and Perceptions of AI

Abstract

This paper investigates connective labor, a novel concept of service work, and examines the impact of contemporary trends in standardization and automation, including artificial intelligence, or AI. Some jobs have relationship at the core of their work, requiring an emotional connection between workers and their charges; teachers, therapists, primary care physicians, each depend on relationship in service to a larger goal: children learning, patients healing. Connective labor is an original term capturing the clinical, relational work between practitioner and recipient, using their emotional connection to produce a particular end.  Existing research documents the importance of work involving relationships for valuable outcomes in arenas from schools to hospitals; these proven impacts of connective labor make its scarcity, uneven distribution or unreliable performance a social problem. Yet its particularistic, emotional nature resists comprehensive systematization or automation. Based on more than 100 interviews with practitioners and recipients as well as engineers, program administrators, and others who systematize that work, this paper 1) considers what we know about what makes connective labor successful, and posits that a crucial component is witnessing work; and 2) evaluates the tradeoffs in its automation, including the costs in providing a thin version of witnessing work versus its greater availability and more predictable performance. As it turns out, at stake are contested notions about the role of shame and vulnerability. The paper concludes by reflecting upon how gender and class intertwine to shape the impact of AI on work, the political implications of the importance of witnessing, and some of the ramifications of automating connective labor. (Paper is longer than the talk and also written in a style meant for a wider audience; the talk would shrink the first three sections, focusing principally on the shame/vulnerability angle and the last section on inequality and politics in its entirety).

Author

• Allison Pugh, University of Virginia