Overwork without an employer: fictitious autonomy and algorithmic control in South Korea's platform economy

초록

Social protection systems in East Asia assume an identifiable employer mediates workers' access to social insurance, occupational safety, and working hour regulation. Platform work upends this assumption. Workers are classified as independent yet remain subordinated to algorithmic workflows, bearing operating costs and risk without corresponding autonomy or protection. This study examines how fictitious autonomy relates to overwork among midnight delivery platform workers in South Korea, and the role algorithmic control plays in that relationship. Using original survey data from 942 workers across multiple platforms, latent class analysis identifies three classes that differ in how contractual independence, process discretion, and cost-bearing combine in practice. The largest class (45.8%), marked by minimal process autonomy alongside a high cost burden, most fully embodies fictitious autonomy. Mediation analysis shows that algorithmic control amplifies the link between this class and overwork, accounting for roughly 20% of the total association, while workers retaining substantive process autonomy resist algorithmic pressure. These findings expose a structural mismatch in employer-centred welfare states. When platforms obscure the employment relationship, regulatory institutions designed to prevent overwork fail to reach the workers who need them most, with implications for social protection reform across East Asia.

키워드