Organizing workers and machine learning tools for a less oppressive workplace
Amber Young, Ann Majchrzak, Gerald C. Kane
Abstract
Machine learning tools are increasingly infiltrating everyday work life with implications for workers. By looking at machine learning tools as part of a sociotechnical system, we explore how machine learning tools enforce oppression of workers. We theorize, normatively, that with reorganizing processes in place, oppressive characteristics could be converted to emancipatory characteristics. Drawing on Paulo Freire’s critical theory of emancipatory pedagogy, we outline similarities between the characteristics Freire saw in oppressive societies and the characteristics of currently designed partnerships between humans and machine learning tools. Freire’s theory offers a way forward in reorganizing humans and machine learning tools in the workplace. Rather than advocating human control or the decoupling of workers and machines, we follow Freire’s theory in proposing four processes for emancipatory organizing of human and machine learning partnership: 1) awakening of a critical consciousness, 2) enabling role freedom, 3) instituting incentives and sanctions for accountability, and 4) identifying alternative emancipatory futures. Theoretical and practical implications of this emancipatory organizing theory are drawn.