Artificial Intelligence and Labour Productivity by Skill Stratification: Empirical Evidence from Chinese Municipal Provinces, 2000–2020

Abstract

Artificial intelligence (AI) is reshaping labor productivity by enhancing production efficiency, transforming work patterns, and redefining labor market dynamics. However, the full potential of AI remains constrained by varying adoption rates and a persistent mismatch between technological advancements and workforce skill levels. Consequently, many regions have yet to realize substantial productivity gains. This study investigates the impact of AI on the productivity of high, medium, and low-skilled workers in four municipal provinces in China from 2000 to 2020. Employing Ordinary Least Squares (OLS) regression on a 21-year panel dataset, the research examines how three proxies of AI adoption, patents, research investments, and infrastructure development affect labor productivity across different skill tiers. The findings reveal significant heterogeneity: AI patents and research investments disproportionately benefit high-skilled workers. At the same time, infrastructure-based AI development is crucial to enhancing productivity for medium- and low-skilled workers. These results underscore the importance of skill-aligned AI strategies to ensure inclusive productivity growth. This study makes an urgent contribution to the discourse on digital transformation and labor market adaptation by offering evidence-based insights for policymakers. It emphasizes the need for coordinated efforts among research institutions, industries, and local governments to promote continuous learning and upskilling. Such collaboration is vital to equip the workforce with capabilities that align with emerging AI technologies, enhancing resilience, competitiveness, and adaptability in rapid technological change.