AI Washing and Stock Returns: Evidence from Chinese A-Share Firms

Abstract

Artificial intelligence has become increasingly prominent in corporate disclosure, yet firms' AI-related narratives may not always be supported by observable investment in technology or human capital. Using AI-related keywords from annual reports, AI patents, and AI-related job postings of Chinese A-share firms from 2020 to 2025, we construct a firm-year measure to examine this disclosure-investment gap. We classify a firm as an AI washer when its AI-related disclosure is above the industry-year median, while both AI patent activity and AI-related recruitment remain below their corresponding medians. Firms meeting this definition are associated with lower annual stock returns, and the result remains qualitatively unchanged when we use alternative methods to identify AI patents. Additional tests point to two related market responses: AI washing is accompanied by more negative media coverage and weaker investor sentiment. The return discount is also more pronounced among firms with higher institutional ownership and among firms in regions with stronger local AI policy support. These results suggest that investors do not treat AI-related disclosure independently of observable corporate activity. Instead, the impact of AI narratives on stock returns appears to depend partly on whether firms can support those narratives with technological development and hiring.