The determinants of AI Washing in Corporate Reporting: Strategic Antecedents and Path Dependence of Symbolic Technological Narratives.

Artificial intelligence has become a prominent element of corporate strategic narratives, yet little is known about what drives firms to overstate their AI orientation and whether such behavior reflects isolated opportunism or persistent organizational routines. We conceptualize AI washing as a symbolic–substantive gap in which firms project an AI-oriented technological identity that is not supported by subsequent verifiable innovation outputs. Building on symbolic management and legitimacy perspectives, we argue that AI washing represents a strategic response to legitimacy pressures and resource incentives under conditions of technological opacity.

Using Chinese A-share listed firms from 2010 to 2024, we develop complementary measures of AI washing that capture the mismatch between AI-related narratives in annual-report MD&A sections and firms’ subsequent AI innovation outputs, combining textual disclosure, R&D intensity, and AI patent applications. We find that financial distress significantly increases AI washing, consistent with firms using future-oriented technological narratives to redirect attention from weak fundamentals. Media visibility and investor attention are also positively associated with AI washing, suggesting that external attention may increase the returns to symbolic technological positioning when underlying technological capabilities are difficult to verify. AI-related government subsidies further increase AI washing, revealing a potential unintended consequence of technology-oriented policy support. Finally, AI washing exhibits strong path dependence: firms with prior histories of greenwashing and digital washing are significantly more likely to engage in subsequent AI washing.

Our findings extend symbolic management research from misrepresentation of realized performance to future-oriented symbolic technological narratives, showing how technological opacity enables firms to construct innovation identities before substantive outcomes become observable. The study also introduces a narrative–output decoupling approach to measuring AI washing and demonstrates that symbolic disclosure practices can persist and transfer across substantive domains.