Fueling sustained green innovation advantage in Chinese listed firms: how AI orientation works through knowledge recombination and knowledge diversity

Purpose Green innovation is critical to long-term advantage, yet sustaining it across multiple dimensions remains difficult. Drawing on the knowledge-based view, attention-based view and dynamic capability theory, this study aims to examine how artificial intelligence (AI) orientation (AIO) promotes sustained green innovation advantage (SGIA) through knowledge mechanisms and under what conditions. Design/methodology/approach Using an unbalanced panel of Chinese A-share listed firms from 2001 to 2023, the authors combine annual-report text, patent and financial data. The authors measure AIO through text mining and validate it with alternative specifications, including an AI-clean measure that removes AI-washing. SGIA is modeled as an intertemporal, multidimensional construct encompassing quantity, quality, efficiency and speed. Main, mediation and conditional indirect effects are estimated using firm fixed effects. Findings AIO significantly improves SGIA and all four dimensions. Knowledge recombination creation (KRC) and knowledge recombination reuse (KRR) mediate these effects, while knowledge diversity (KD) more robustly strengthens the indirect pathway via KRR than via KRC. Originality/value This study conceptualizes SGIA as a multidimensional, intertemporal construct, distinguishes substantive AIO from symbolic AI narratives through AIW and AI-clean measures and identifies KRC and KRR as key mechanisms and KD as a selective boundary condition linking AIO to SGIA.