Can Artificial Intelligence Reduce the Carbon Intensity of Manufacturing Industry?—A Study Based on Spatial Threshold Panel Model

探讨人工智能对制造业碳强度的影响对于推动传统制造业向绿色、智能、高端转型升级具有重要的现实意义。基于2006~2022年中国省域面板数据，采用空间门限面板模型考察了中国省域制造业碳强度的空间溢出效应以及人工智能对制造业碳强度的非线性门槛效应。研究结果表明：(1) 中国制造业碳强度存在正向空间相关性，呈现出“高–高”和“低–低”的集聚特征。(2) 相邻省份制造业碳强度每增加1个单位，观测省份的制造业碳强度提高0.182个单位。(3) 人工智能对制造业碳强度的影响在经济发展水平下存在两区制门槛效应。当PGDP > 18768.78元时，人工智能对制造业碳强度的降低具有积极作用，反之，其对制造业碳强度的影响则不显著。根据上述结论，本文提出相应的政策建议。Exploring the impact of artificial intelligence on the carbon intensity of manufacturing industry has important practical significance for promoting the transformation and upgrading of traditional manufacturing industry to green, intelligent and high-end. Based on China’s provincial panel data from 2006 to 2022, a spatial threshold panel model was used to investigate the spatial spillover effect and the nonlinear threshold effect of artificial intelligence on carbon intensity in China’s provincial manufacturing industry. The results show that: (1) The carbon intensity of China’s manufacturing industry has a positive spatial correlation, showing “high-high” and “low-low” agglomeration characteristics. (2) For every 1 unit increase in manufacturing carbon intensity in neighboring provinces, the observed province’s manufacturing carbon intensity increases by 0.182 units. (3) The influence of artificial intelligence on the carbon intensity of manufacturing industry has a two-zone threshold effect at the level of economic development. When PGDP is greater than 18768.78 yuan, artificial intelligence has a positive effect on the reduction of carbon intensity in the manufacturing industry, on the contrary, its impact on the carbon intensity of the manufacturing industry is not significant. According to the above conclusions, this paper puts forward corresponding policy recommendations.