Speculation in the Age of AI: Measuring AI Wash and Its Valuation Consequences

Description

The escalating discourse on artificial intelligence (AI) has intensified investors' and financial regulators' demands for assessing the authenticity of AI-related disclosures. This paper proposes a novel ranking-based discrepancy measure to quantify AI wash—the divergence between firms' AI-related rhetoric (measured through keyword prominence in earnings calls) and their substantive AI efforts. The results suggest that investors initially overreact to AI wash, leading to short-term overvaluation in the first three quarters following AI wash disclosure. However, this effect reverses over the subsequent seven quarters, suggesting that AI wash ultimately results in long-term valuation corrections. Further analysis reveals that these reversals are more pronounced for firms with lower institutional ownership and higher information asymmetry. Collectively, our results indicate that AI-related disclosures are subject to informational frictions, highlighting the need for greater scrutiny and potential regulatory oversight.

Period	10 Nov 2025
Held at	Zayed University, United Arab Emirates
Degree of Recognition	International