Events

(joint work with Daniel Stegmueller)

Abstract: While the organization of workers in labor unions has underpinned shared prosperity and dampened inequalities in political representation, recent research shows that advances in automation based on robotization and artificial intelligence (AI) weaken labor unions. Thus, automation that displaces unionized jobs may nip in the bud the post-Covid-19 revitalization of workplace organization and end democracy at work as we know it. While most of the recent research points in one direction, technological change need not be uniformly bad for worker bargaining power and voice. We argue that sometimes automation can empower workers to organize by, first, increasing their bargaining power and mobilization potential and, second, broadening coalitions in favor of democracy at work. We employ a simple model of workplace elections to illustrate the bargaining mechanism. It highlights that investments in automation and public policies that reduce firms’ threat of capital flight should increase workers’ incentives to support unions. In study 1, we test implications of the argument in an original vignette experiment embedded in a survey of 2,500 US workers. We find that workers are receptive to arguments that highlight the prospects for unionization in the face of firm opposition and future automation. We discuss a follow-up design (study 2). To investigate the possibility that AI-driven automation increases the demand for democracy at work — particularly among a broader range of occupations including the college-educated — we propose a choice experiment that examines whether firms' AI investments increase workers' demand for unions (study 3).

The Political Science Seminar Series is convened by Desmond King and David Rueda For more information on this or any of the seminars in the series, please contact politics.secretary@nuffield.ox.ac.uk.