中国地质大学学报(社会科学版）

随着“双碳”战略目标的深入推进，碳减排已成为实现可持续发展的核心引擎。人工智能作为引领新一轮科技革命和产业变革的战略性技术，成为驱动绿色低碳转型的重要力量。本文基于中国284个城市面板数据，采用固定效应模型、中介效应模型、门槛效应模型等方法，多维度实证考察了人工智能对碳减排的影响效应。研究发现：（1）人工智能显著降低了碳排放强度，该效应在经过一系列稳健性检验后依然成立；（2）绿色技术创新和产业结构优化的“增量”和“提质”作用是人工智能有效降低碳排放强度的重要机制；（3）在东部地区、中心城市、数字化水平较高以及环境规制较高的城市，人工智能对碳减排的促进效应更强；（4）人工智能对碳减排呈现出边际效益先递增后递减的非线性特征，过度关注人工智能可能引发潜在的风险。

As the “dual carbon” strategic goal advances, carbon emission reduction has become the core engine for achieving sustainable development. Artificial intelligence, a strategic technology leading the new round of scientific and technological revolution and industrial transformation, has become a vital force driving the green and low-carbon transition. Based on panel data from 284 Chinese cities, this paper employs fixed effects models, mediation effect models, and threshold effect models to conduct a multi-dimensional empirical examination of the impact of AI on carbon emission reduction. The findings reveal that:(1) Artificial intelligence significantly reduces carbon emission intensity, a result that remains robust after a series of sensitivity tests;(2) The “incremental” and “quality improvement” roles of green technological innovation and industrial structure optimization are key mechanisms through which artificial intelligence effectively lowers carbon emission intensity;(3) The promoting effect of artificial intelligence on carbon emission reduction is stronger in eastern regions, central cities, cities with higher digitalization levels, and cities with stricter environmental regulations;(4) Artificial intelligence exhibits a nonlinear feature of marginal benefits, first increasing and then decreasing, suggesting that excessive focus on artificial intelligence may trigger potential risks.