Against the backdrop of the diffusion of generative artificial intelligence (GenAI) and the expanding demand for external green technology under the dual-carbon goal, this paper aims to examine how firms’ levels of GenAI adoption affect their acquisition of green technological knowledge from external sources, identify two transmission mechanisms and characterize the firm-level boundary conditions that shape the marginal value of GenAI.
Using panel data on listed firms from 2014 to 2023, the authors measure firm-level GenAI by the log-transformed frequency of generative-AI-related keywords in annual reports and measure green knowledge transfer by the log-transformed number of green patents received from external assignors.
Firms’ GenAI levels exert a robust positive effect on their external green knowledge transfer, and instrumental-variable estimates survive weak-instrument and over-identification tests with a causal interpretation. GenAI operates through two independent channels: enhancing access to cross-organizational collaboration and raising R&D intensity. Although the positive moderating effect of internalization capability is directionally consistent, it is not statistically significant, while its quadratic term reveals diminishing marginal returns. Heterogeneity analysis reveals four conditional dimensions – substantive adoption, management orientation, supply-chain resilience and green finance pilot status – along which directional between-group differences emerge. Firms that disclose AI keywords without substantive investment display a reverse co-movement with green knowledge transfer.
This paper links the literature on general-purpose technologies with that on inter-organizational knowledge transfer by proposing and testing a mechanism through which GenAI facilitates green knowledge flows, specifically by reducing identification and evaluation barriers in cross-organizational collaboration.
