AI-Oriented Digital Governance and Financial Transparency: Evidence from Chinese Listed Firms

This study investigates whether Chinese listed firms’ financial transparency improves with AI. Our main AI measure is AI-related disclosure intensity based on annual report text. We validate it using firm-year proxies for AI-related patents and AI/software investment to reduce potential symbolism. We add extensive controls, governance covariates, and mechanism tests to support the instrumental-variable results. From 2011 to 2023, we measured transparency through stock-exchange disclosure quality ratings and AI through annual report AI-related language for 1,028 Shanghai and Shenzhen Stock Exchange firms. Using provincial internet penetration as an AI instrument, we estimate two-way fixed effects models and address endogeneity using two-stage least squares. We also compare the effect by firm size, ownership (SOE vs. non-SOE), industry (high tech vs. non-high-tech), and regional development. Even after controlling for firm characteristics, absorbing firm and year effects, and instrumenting AI, firms with higher AI intensity are more transparent. In smaller, privately owned, and high-tech firms, and in firms located in less-developed regions, transparency gains are greatest. The findings suggest that AI-supported monitoring and documentation improve reported information and can partially replace poorer internal control capabilities.