Hanzhe Li (University of Hong Kong); Jin Li (University of Hong Kong); Andrew Yunchou Zhang (University of Hong Kong)
Abstract
We study when firms replace human managers with AI. We develop a relational contracting model in which a principal hires a worker on a fixed wage and delegates supervision to either a human or an AI manager. The worker's effort cost fluctuates over time and is privately observed by the worker. Human managers can use real-time, context-specific information to tailor supervision to the worker's state. AI managers condition only on public information but have a commitment advantage in enforcement. We characterize the firm's optimal choice of the manager. In a benchmark where the firm can design any positive fixed wage, higher future surplus (greater patience) makes human management more attractive. We then study minimum wages. A higher minimum wage increases AI adoption and can lower worker welfare by inducing a shift toward stricter, less flexible supervision. Minimum wages also reshape matching inside firms. When the wage floor is absent or non-binding, higher worker ability increases AI adoption. When the wage floor binds, higher ability instead increases the use of human managers. These results link AI adoption in management to labor-market institutions and worker heterogeneity, with implications for productivity, welfare, and organizational design.