Artificial intelligence, green finance and urban energy efficiency: Evidence from Chinese 282 cities

Rapid improvements in urban energy efficiency (UEE) are essential for achieving climate and sustainable development goals, yet the roles of artificial intelligence (AI) and green finance in this process remain insufficiently understood. This study develops a theoretical model that links AI to UEE through technological innovation and industrial structure adjustment, and examines the role of green finance. Then, using panel data for 282 Chinese cities from 2012 to 2023, we conduct an empirical analysis to tests the theoretical framework. The main findings are as follows. (1) AI significantly improves UEE and this finding holds following a series of robustness and endogeneity tests. The positive effect is not universal but is primarily observed in the cities with greater location, industry conditions, and government attention. (2) Green technological innovation as well as the rationalization and advancement industrial structure are key channels through which AI improves UEE. (3) Green finance amplifies the benefits of AI by easing financing constraints, and exhibits a nonlinear threshold effect whereby the marginal contribution of AI to UEE increases once green finance exceeds a critical level. (4) Further analysis reveals that AI exhibits positive spatial spillovers, does not induce an energy rebound effect, and reduces urban carbon emission intensity. We also found that human-machine collaboration plays a crucial role on UEE. This study provides theoretical and empirical evidence for policymakers to develop AI and energy strategies in city level.