Drivers of innovation in China's AI industry: A systematic review

This systematic review summarizes firm-level evidence on factors influencing innovation performance in China's AI industry. A PRISMA-style search in Web of Science and Scopus identified 38 empirical studies. Innovation outcomes are mainly measured with patents (n=15) and scales or indices (n=10). Other outcomes include green innovation (n=3), product or new-product outcomes (n=2), and mixed or other outcomes (n=8). Across studies, AI adoption and related digital inputs rarely show apparent direct effects on innovation. Most evidence supports an indirect path through capability building. Key factors are organizational learning, knowledge management, human capital, and social capital. Many studies test these factors as mediators, often through an innovation-capability layer. Effects also depend on China-specific conditions, including major AI hubs and regional gaps, policy intensity and design, capital cycles and platform ecosystems, and limits in talent, compute, and data quality. Based on these findings, the review proposes a testable framework and identifies gaps in measurement, research design, and multi-level linkage.