Artificial Intelligence and Intergenerational Income Mobility in China: Evidence from GPU Imports

• In Event: Inequalities in Income, Wealth, and Related Outcomes

Abstract

This study investigates the impact of Artificial Intelligence (AI) development on intergenerational income mobility in China, a context characterized by rapid technological adoption and deeply entrenched social structures. Integrating 2010–2022 China Family Panel Studies (CFPS) microdata with provincial-level high-performance Graphics Processing Units (GPUs) import data, we employ a moderated regression framework with Weighted Ordinary Least Squares (WOLS) model to examine whether AI functions as a "Great Equalizer" or an "Entrencher of Privilege." Our empirical analyses robustly support the Reinforcing Hypothesis that AI development significantly perpetuates intergenerational income persistence, crystallizing pre-existing economic advantages. Specifically, each log-point rise in provincial GPU imports amplifies the parental-income coefficient on children’s earnings (or ranking) by approximately 0.01, thereby hardening the intergenerational income nexus. Heterogeneity analysis reveals a stark generational divergence; this reinforcing effect is concentrated exclusively among the post-reform cohort (born after 1980s), whose educational and early-career stages coincided with rapid digitalization process after 2000. No significant effect is detected for earlier cohorts. These findings remain robust across alternative AI proxies and rank-rank specifications. Our results suggest that without targeted policy interventions to equalize digital human capital and reform institutional barriers, AI may act as a catalyst for social stratification rather than a vehicle for inclusive growth.
Keywords: Artificial Intelligence, Intergenerational Mobility, Social Stratification, Quantitative Social Science, China

Authors

• Junru Chen, Department of Sociology

• Linfeng SHEN, The Chinese University of Hong Kong