From innovation to uncertainty: artificial intelligence narratives and financial market volatility

Abstract

The paper explores the relationship between artificial intelligence (AI)-related uncertainty and financial market volatility. We isolate the uncertainty component of AI narratives using indexes from newspaper-based information from 2015 to 2025, which allows us to differentiate between general AI attention, AI-related economic discourse, and AI-related economic uncertainty. We observe that uncertainty-oriented AI narratives have gained momentum since the advent of generative AI and are most pertinent to market volatility. While general AI coverage and economic discussion do not contain predictive information for the CBOE Volatility Index (VIX), predictive regressions indicate that uncertainty around AI does contain predictive information. The findings are consistent across lag structures, normalization methods, and sub-period analyses, though they are somewhat weakened when VIX persistence is explicitly modelled. The relationship is dynamic: the association becomes significantly weaker during the emergence and diffusion of generative AI. During the pre-ChatGPT period, a one-standard-deviation increase in AI-related economic uncertainty is associated with an approximately 7.08-point increase in VIX, whereas during the post-ChatGPT period, this relationship is significantly weakened. Overall, the findings show that economic uncertainty surrounding AI is a unique narrative-based indicator of market uncertainty and that its economic salience relies on the framing of technological developments.