Synergistic Effects of AI Pilot Zones and Climate Governance on Urban Green Transformation: Evidence for Sustainable Urban Development

Abstract

This study is a dedicated empirical investigation of the impacts of Artificial Intelligence Pilot Zones on sustainable urban development, focusing on the ecological and technological dimensions. Coordinate the promotion of digital technology empowerment and climate adaptation to advance the green economic decarbonisation of developing countries. Although previous studies have focused on digital pilots and environmental regulations separately, the spatial synergy and micro-level transmission mechanisms of technological instruments under institutional constraints have not been explored. Leveraging the spatial overlap between the National Artificial Intelligence Innovation and Development Pilot Zones and Climate-Adaptive City Pilots as a quasi-natural experiment, this paper uses a balanced panel dataset of 285 Chinese prefecture-level cities from 2000 to 2022. Based on Porter's hypothesis and the theory of technology-institution complementarity, we construct a Constraint Instrument theoretical framework and apply multi-period interaction difference-in-differences (DID), Callaway and Sant'Anna (2021) heterogeneity-robust estimators, and partially linear double machine learning (DML) to identify policy synergy. The empirical results show that policy synergy significantly reduces the intensity of urban industrial SO2 emissions and energy-intensity per unit of GDP, with an interaction coefficient of -0.0542 (p < 0.05), indicating a complementary "1 + 1 > 2" decarbonisation dividend. The dynamic parallel trend assumption is verified over an extended pre-treatment window (τ = -5 to +3) using a joint Wald test (p = 0.702) anchored on joint policy rollout, and anticipation effects have been ruled out. The results are consistent after controlling for the concurrent policies (Low-Carbon Cities, Smart Cities, Sponge Cities), meteorological factors (temperature and precipitation), and 500 Monte Carlo placebo runs. Disentangled mechanism analysis shows that green invention patenting (accounting for 28.6% of SO2 reduction) and energy-saving improvements (accounting for 34.3%) are the main transmission paths, while green digital process innovation and industrial rationalisation drive energy-intensity reduction. Heterogeneity analysis shows that the synergistic dividend is significantly higher in old industrial cities (-0.0782), and high computing-density hubs show a slight, statistically non-significant decrease, indicating a computing-energy rebound effect. The above results provide empirical support and a transferable model for the development of a coordinated digital transformation and climate adaptation strategy in China. From the dual perspectives of technological innovation and ecological emission abatement, this study provides granular empirical evidence for the role of AI pilot zones in driving sustainable urban development.