Enabling or weakening? Artificial intelligence and ambidextrous green innovation

Purpose This study examines how artificial intelligence (AI) adoption shapes firms' ambidextrous green innovation by inducing differential effects on exploratory and exploitative activities from an attention-based perspective. Design/methodology/approach Drawing on the attention-based view, we theorize that AI adoption reconfigures organizational attention and technological capabilities, leading to divergent effects on green innovation. Using panel data from Chinese listed firms, we employ regression analyses to test these arguments. Findings AI adoption exerts a double-edged effect on ambidextrous green innovation. It significantly promotes exploratory green innovation by redirecting managerial attention toward forward-looking strategic issues, while suppressing exploitative green innovation by institutionalizing routine processes and reducing sustained attention to incremental improvements. Furthermore, Confucian cultural embeddedness, negative media coverage and institutional ownership strengthen the positive effect of AI on exploratory green innovation and mitigate its negative effect on exploitative green innovation. Practical implications These findings suggest that managers should align AI adoption with balanced attention allocation to avoid overemphasizing exploratory innovation at the expense of incremental improvements. Originality/value This study advances the literature by conceptualizing AI as an attention-restructuring mechanism that systematically differentiates firms' exploratory and exploitative green innovation. It contributes to research on AI and organizational ambidexterity by uncovering the contingent role of institutional and social contexts.