Occupational AI Substitution Risk, Skill Mismatch, and Individual Income Inequality: Evidence from China

This study examines how occupational exposure to artificial intelligence substitution risk is associated with individual income inequality in China. Using China Family Panel Studies data matched with occupation-level AI substitution risk and O*NET skill-demand information, we decompose income inequality into overall, within-occupation, and between-occupation components based on Kakwani’s relative deprivation framework. The results reveal a dual distributional effect. AI substitution risk is negatively associated with within-occupation inequality, suggesting that AI-related task standardization may compress income differences among workers in similar occupations. However, it is positively associated with between-occupation inequality and overall inequality, indicating that AI substitution risk may widen income gaps across occupational groups. Mechanism tests show that AI substitution risk is associated with lower skill shortage but higher skill surplus, suggesting that skill adjustment under technological pressure may generate both better within-occupation matching and broader human-capital underutilization. Heterogeneity analysis further shows that older workers and flexible workers are more vulnerable to the inequality-enhancing effects of AI substitution risk. These findings suggest that AI does not reshape income inequality in a uniform direction, but reorganizes labor-market inequality through occupational structure, skill mismatch, and worker vulnerability.