A study on the impact of urban artificial intelligence development on carbon emissions from electricity consumption

The advancement of artificial intelligence in conjunction with the “dual carbon” goals is a critical aspect of the current green transition of cities. The carbon reduction effects of this approach in the electricity consumption sector require a thorough evaluation. Using panel data from 287 prefecture-level and higher Chinese cities from 2010 to 2023, this paper empirically examines how urban artificial intelligence development impacts carbon emissions from electricity consumption and the underlying mechanisms. The research found that: (1) The development of artificial intelligence in cities can significantly reduce carbon emissions from electricity consumption. This conclusion remains valid after a series of robustness and endogeneity checks. Second, analysis of the mechanism of action indicates that urban artificial intelligence development primarily curbs carbon emissions from electricity consumption through three channels: enhancing government regulation, promoting industrial restructuring, and stimulating green technological innovation among enterprises. Third, heterogeneity analysis shows that the carbon emission reduction effects of AI development on electricity consumption are more pronounced in cities located northwest of the “Hu-Huan-Yong Line,” in peripheral cities, and in resource-based cities. This reflects the fact that AI has a greater marginal impact on reducing emissions in regions with significant potential for improving energy efficiency and with an industrial structure heavily weighted toward resource-intensive sectors. This research provides policy guidance for promoting the deep integration of AI with green, low-carbon development and for exploring tailored carbon reduction pathways for cities.