Le Yan, Wei Li, Shizheng Tan, Xiaoguang Liu
Using panel data for 30 provinces in mainland China (2013–2022), this research examines how artificial intelligence (AI) affects green innovation resilience (GIR) and the mechanisms through which this occurs. It tests industrial structure advancement and industrial structure rationalization as mediating channels, and evaluates threshold effects associated with public environmental concern and environmental regulation. The results indicate that AI is positively and significantly related to GIR, and the conclusion remains stable under multiple alternative specifications and robustness checks. Further analysis reveals that different dimensions of industrial structure upgrading play distinct roles. AI indirectly strengthens innovation resilience through the quantity and quality dimensions of industrial structure advancement, whereas industrial structure rationalization does not constitute an effective transmission channel, highlighting heterogeneity in technological–structural synergy. Moreover, the threshold effects of public environmental concern and environmental regulation differ markedly. Public environmental concern exhibits a critical threshold that needs to be maintained within a reasonable range, whereas stronger environmental regulation amplifies the technological dividends of AI in a staircase reinforcement pattern. Overall, this study systematically explores the mechanisms and boundary conditions through which AI drives green innovation resilience, providing new theoretical insights and empirical evidence for green transformation in the AI era.