In the context of the rapid advancement of artificial intelligence (AI), the internal mechanisms through which AI applications affect green productivity within firms remain inadequately understood. This study empirically investigates the relationship between AI application levels and green total factor productivity (GTFP) among Chinese listed firms from 2007 to 2023, yielding the following conclusions. (1) A significant inverted U-shaped relationship exists between the level of AI application and firms' GTFP, which remains robust after a series of stringent tests. (2) Mediation analysis reveals that AI applications exert an inverted U-shaped influence on GTFP by initially enhancing, then subsequently suppressing the degree of green transformation of strategy management and the efficiency of green technology R&D. (3) Regarding moderation effects, both the regional higher education level and a firm's corporate governance quality attenuate the inverted U-shaped relationship, leading to a flattening of the curve. (4) Heterogeneity analysis indicates that the inverted U-shaped relationship is more pronounced in firms facing low environmental regulation, low emissions, and high financing constraints. The flattening effect of corporate governance is more substantial in firms with low environmental regulation, low emissions, and low financing constraints. In contrast, regional higher education exerts a more significant moderating role in firms subject to high environmental regulation, with low emissions and low financing constraints. This paper contributes to understanding how the intra-firm application of AI influences the development of green productivity through strategic change and technological innovation, and provides a novel perspective for reconciling their divergent macro-level impact mechanisms. • Decode AI–GTFP nonlinear nexus inside firms. • Managerial & tech lenses explain AI’s nonlinear green effect. • Governance & external talent moderate AI–GTFP slope. • Micro AI green gains scale to dual macro channels.
Building similarity graph...
Analyzing shared references across papers
Loading...
Ying et al. (Sat,) studied this question.
www.synapsesocial.com/papers/69a75f76c6e9836116a2adaf — DOI: https://doi.org/10.1016/j.iref.2026.104981
Ke Ying
Luo Jiajun
International Review of Economics & Finance
Center for Strategic Research
University of Economics and Innovation
Building similarity graph...
Analyzing shared references across papers
Loading...