Artificial intelligence innovation and environmental performance: Unraveling the complex roles of application and method innovation across enterprise sizes

This study examines the complex impact of artificial intelligence innovation (AII) on enterprise environmental performance, addressing a critical gap in understanding its dual effects. Grounded in the Knowledge-Based View, we differentiate between AI application innovation and AI method innovation, analyzing their distinct environmental outcomes using data from Chinese-listed enterprises (2016–2021). Our findings reveal that AI application innovation follows a U-shaped relationship with environmental performance: early stages may lead to increased resource consumption and pollution, while positive effects emerge as adoption scales up. In contrast, AI method innovation consistently enhances environmental performance. The study further identifies the moderating role of enterprise size, showing that larger enterprises experience stronger positive effects from AI method innovation, while the U-shaped relationship of AI application innovation becomes flatter as enterprise size increases. These insights provide a nuanced understanding of AII's environmental implications, contributing to the literature by clarifying both linear and nonlinear effects. The findings offer practical guidance for enterprises to optimize AI strategies and advise policymakers on tailoring support measures to promote sustainable AI adoption across various organizational contexts.