Tian tian Mu, Shan Qiaojuan, Zhu Guangyu, Xu Yang, Gomaa Agag, Osman Alipour
This research explores the dual role of artificial intelligence (AI) and its narratives in shaping environmental, social, and governance (ESG) performance in China's retail industry. Drawing on the dynamic capabilities view and signaling theory, we conceptualize AI as both a substantive technological resource and a symbolic communicative mechanism, with green innovation serving as a mediating process. Using panel data of 2306 A-share listed retail firms from 2020 to 2024 (9710 firm-year observations), we employ system generalized method of moments (GMM) estimations to address dynamic panel bias, endogeneity, and unobserved heterogeneity. The findings reveal that both AI technology and AI narrative have a positive impact on ESG performance, with AI technology exerting a stronger substantive effect and AI narratives functioning as credible signals that enhance legitimacy. Moreover, green innovation significantly improves ESG outcomes and partially mediates the artificial intelligence - ESG performance relationship. Robustness checks using alternative measures, estimators, subsamples, and placebo tests confirm the stability of the findings. This research contributes to the literature by linking digital transformation and sustainability, highlighting the interplay between substantive and symbolic mechanisms, and providing insights from an emerging economy context. The findings offer theoretical, managerial, and policy implications for leveraging AI to advance sustainable development in retailing.