Generative AI and responsible supply chains: When sustainability gains undermine social legitimacy

Generative artificial intelligence (GenAI) is increasingly adopted in supply chains to advance sustainability goals; however, empirical evidence on its environmental and social consequences remains limited and fragmented. This study examines how GenAI adoption affects responsible supply chain outcomes by integrating Difference-in-Differences (DiD) and fuzzy-set Qualitative Comparative Analysis (fsQCA). Using firm–quarter panel data from 320 South Korean firms covering the period from 2023Q1 to 2025Q2, we estimate the short-run causal effects of GenAI adoption on environmental, operational, and social performance, while also identifying the organizational configurations under which these effects emerge. The DiD results provide causal evidence that GenAI adoption significantly reduces CO2 intensity and strengthens circular supply chain practices in the short term. Mediation analyses further show that these environmental benefits are realized primarily through enhanced circular practices rather than through direct improvements in operational efficiency. In contrast, the average effects on operational efficiency and social responsibility outcomes are statistically insignificant, indicating substantial heterogeneity across firms. The fsQCA analysis uncovers multiple sufficient pathways to sustainability. GenAI adoption combined with high absorptive capacity consistently produces environmental gains, whereas the combination of GenAI, responsible AI governance, and strong managerial commitment is the only configuration that generates joint environmental and social outcomes. Conversely, GenAI adoption in the absence of governance safeguards leads to environmental improvements while undermining social legitimacy, underscoring the conditional and double-edged nature of GenAI in supply chains. By integrating causal inference with configurational analysis, this study advances understanding of when and how GenAI contributes to responsible supply chains. The findings demonstrate that while GenAI can serve as a powerful enabler of circular and low-carbon supply chains, its broader social legitimacy critically depends on complementary organizational capabilities and governance arrangements.