Artificial intelligence in urban land use: How regional policy and institutional embeddedness shape the economic efficiency-social legitimacy paradox

Urban land use policy increasingly intersects with digital transformation strategies as cities worldwide pursue smart development agendas. This study examines how regional institutional frameworks—encompassing land use regulations, digital economy policies, and urban development priorities—shape the adoption and outcomes of artificial intelligence technologies among urban land users. Focusing on the hospitality sector as intensive urban land occupants operating under diverse regulatory regimes, we investigate how firms navigate institutional pressures while deploying AI-enabled service systems. Using structural equation modeling with data from 587 hotels across China, Europe, and the United States, complemented by 33 executive interviews, we reveal that regional institutional embeddedness creates paradoxical effects: while enhancing social legitimacy and stakeholder acceptance of technological change, it simultaneously constrains operational efficiency gains through conformity costs. Digital employee integration exhibits convex, J-curve-like trajectories within the observed data range for both efficiency and legitimacy outcomes, challenging linear assumptions about technology-driven land use optimization. These findings demonstrate that effective urban land use policy must account for the temporal dynamics and institutional trade-offs inherent in digital transformation, recognizing that rigid institutional alignment may impede the operational flexibility necessary for urban land users to maximize spatial productivity. The research offers policy guidance for governments seeking to balance innovation encouragement with social acceptance in urban development contexts, while revealing how place-based institutional strategies shape the economic geography of digital adoption. • Regional embeddedness creates an efficiency-legitimacy paradox for firms. • AI integration mediates strategy to optimize urban land use performance. • Adoption outcomes follow J-curve-like trajectories in urban service sectors. • Rigid policy alignment constrains operational flexibility in land use. • Study of 587 firms reveals how place-based institutions shape AI adoption.