How AI reduces pollutant emissions: The dual mechanisms of productivity enhancement and financing constraint alleviation

Abstract

Artificial intelligence (AI) technology is an indispensable driving force in promoting the green transformation of enterprises and high-quality development. Based on China's A-share listed companies, this study establishes firm-level AI technology indicators constructed from annual reports via semantic extraction based on GLM-4 large language model to empirically examine the impact of AI on corporate pollution behavior and its mechanism. The research findings indicate that AI technology can effectively reduce the intensity of corporate pollutant emissions. And this conclusion is robust to various tests. Mechanistically, AI improves the TFP of firms and alleviates financing constraints, which helps firms achieve the emission reduction target. Heterogeneity analysis further shows that the emission-reduction effects of AI are much more significant in non-state-owned enterprises, non-heavy-polluting industries, and in firms in sectors with lower competitive intensity. The study findings provide useful references for using AI technology to promote corporate environmental governance and green transformation, and also provide important references for policymakers to formulate differentiated policies according to the characteristics of various enterprises and industries.