Evaluating the pollution abatement effect of artificial intelligence policy: Evidence from a quasi-natural experiment in China

Artificial intelligence (AI) policy refers to a series of regulations, strategies and measures formulated by governments, international organizations or industry bodies to guide the research, development, deployment and application of AI. In recent years, the potential of AI policy in pollution abatement has received widespread attention. In this study, we use China's National New Generation Artificial Intelligence Innovation and Development Pilot Zone (AIIDPZ) policy as a quasi-natural experiment to evaluate the impact of AI policy on urban environmental pollution (EP). Specifically, we construct a staggered difference-in-differences (DID) model and find that the AIIDPZ policy help reduce EP. In terms of potential mechanisms, we find that the AIIDPZ policy mainly abate EP through increasing fiscal technology expenditure, improving the level of green technology innovation and promoting economic agglomeration. Heterogeneity test reveals that the pollution abatement effect is more significant in cities with high levels of EP, as well as eastern cities, southern cities, non-resource-based cities and gigabit cities. Our findings provide important references for policymakers to scientifically utilize AI policies to achieve sustainable development.