Abstract: The impact of corporate activities on air pollution is a crucial aspect of ESG evaluation. In the digital era, where firms increasingly adopt artificial intelligence (AI) technology to drive production transformation, a nuanced understanding of the relationship between AI and air pollution is essential for accurately assessing corporate ESG performance. This study uses panel data from 256 Chinese cities between 2003 and 2020 to examine the impact of AI development on air pollution. We find that AI development, measured by AI patents, increases PM2.5 emissions. This conclusion remains robust across a series of tests, including alternative measurements of the independent variable, exclusion of policy interference, removal of special samples, and addressing endogeneity concerns with instrumental variables. Heterogeneity analysis reveals that utility model AI patents primarily drive the increase in air pollution, with the impact of AI development on air pollution levels being more pronounced in small cities, non-core cities, and cities with weaker air pollution control efforts. Regarding underlying mechanisms, AI development exacerbates air pollution through increased energy consumption and expanded industrial output. Our study underscores the necessity of including AI-driven air pollution externalities in assessing corporate ESG performance.

Keywords: Artificial Intelligence; Air Pollution; Energy Consumption; Industrial Output; ESG

JEL codes: M14; O33; Q53

DOI: ...