Generative Artificial Intelligence, Disclosure Quality and Capital Flows: Evidence From China's Mutual Fund Market

ABSTRACT

We examine how the adoption of generative artificial intelligence (GAI) in fund disclosures affects capital flows in the mutual fund industry. Using 19,029 periodic reports issued by Chinese mutual funds, we identify AI‐generated content with a validated detection algorithm and study its implications for investor behaviour. We find that funds using GAI in disclosure narratives attract significantly higher subsequent fund flows. Mechanism analyses show that AI‐assisted reports are more readable, exhibit greater linguistic certainty, and contain more accurate information, thereby reducing investors' information frictions. These effects are stronger for smaller and younger funds, and are more pronounced among low‐performance funds, consistent with weaker flow‐performance sensitivity. Additional validation, placebo tests, propensity score matching(PSM), and subsample analyses of small and new fund families further confirm the robustness of our findings. Overall, our evidence suggests that GAI is an important informational technology shaping disclosure quality and investor capital allocation.