Artificial intelligence diffusion, educational capacity, and income inequality in China: evidence from structural transmission channels

Abstract

Abstract The rapid diffusion of artificial intelligence (AI) is transforming economic structures and labor markets, yet its implications for income distribution remain unclear in economies characterized by regional disparities, urban–rural segmentation, and unequal educational capacity. Using Chinese provincial panel data from 2010 to 2024, this study examines how AI diffusion is associated with income inequality and whether education weakens this structural transmission. We construct a composite AI diffusion index and validate it with alternative measures, including robot penetration, AI-related firms, and AI patents. Fixed-effects and dynamic panel system GMM estimates show that AI diffusion is positively associated with both the Theil index and the urban–rural income gap. Channel analysis reveals asymmetric mechanisms: AI induces directional labor reallocation, amplifies inequality through industrial structural deviation, and weakens rural industrial integration, an otherwise inequality-reducing channel. Educational capacity mitigates part of the AI–inequality relationship, but its effects are heterogeneous. Basic education supports broad adaptability, higher education helps reduce industrial structural deviation, and vocational education facilitates labor adjustment and rural integration. Educational quality, especially rural quality and the urban–rural quality gap, further conditions education-based blocking effects. The findings highlight educational capacity for inclusive AI diffusion in China.