CONCEPTUALIZING PRECISION LABOR IN ARTIFICIAL INTELLIGENCE TRAINING

Abstract

Accuracy and precision are among the central values in the ML communities and tech industry. What does it take to achieve a high level of technical accuracy? What are the harms resulting from technology companies' obsession with technical accuracy and precision, and who incurs the greatest burdens? This paper explores accuracy in the context of AI training in China. Drawing on 9-month multi-sited ethnographic fieldwork, we document workers’ everyday working practices and challenges and harms under the guise of achieving extreme levels of technical precision demanded by the clients and ML practitioners. We introduce the notion of precision labor, referring to the hidden work involved in erasing the messy, ambiguous, and uncertain aspects of technology production, all in the pursuit of presenting technology as objective, truthful, and high-quality. This notion provides a lens to understand the disproportionate impact of unnecessary and unrecognized labor on digital labor communities within AI production and the emerging harms on them, such as financial precarity and machine subordination. It joins existing work on the prevailing values in ML communities, questions the legitimacy and sustainability of the pursuit of performative accuracy, and calls for enhanced reflexivity and timely intervention.