Technological empowerment: the governance effect of bank artificial intelligence adoption on corporate shadow banking

Using loan-level matched data between banks and firms from 2014 to 2022, we examine the association between banks’ adoption of artificial intelligence and shadow banking activities among non-financial firms. We find that bank-level AI adoption is associated with a significant and robust reduction in firms’ engagement in shadow banking. This mitigating effect operates through two channels. First, AI adoption enhances banks’ ex-ante screening capacity, improves borrower identification, expands credit access for high-quality firms, and alleviates financing constraints. Second, by strengthening post-lending monitoring, AI adoption is associated with a lower incidence of corporate financial misconduct and limits firms’ ability to divert loan proceeds into shadow banking activities. The governance effect is more pronounced among firms with greater information asymmetry and in regions with weaker financial regulation, and it is strongest for entrusted loans, a major form of corporate shadow banking in China. Further evidence shows that bank AI adoption is associated with reduced firms’ reliance on short-term borrowing and significantly higher investment in core business activities, suggesting that AI curbs financial arbitrage incentives and facilitates the reallocation of credit towards the real economy. This study provides novel bank–firm-level evidence on how digital financial intermediation reshapes corporate shadow banking behaviour.