spb/uqo-research-db

1% ============================================================================2\section{Conclusion}3\label{sec:conclusion}4% ============================================================================56Using a semantic, embedding-based measure of artificial-intelligence language in the SEC filings of7S\&P~500 firms, linked to Yahoo Finance returns and standard factor data, we have shown that the8pricing of AI disclosure is a \emph{conditional, regime-dependent} phenomenon. For most of the 2010s,9how intensively a firm wrote about adopting, building, or being threatened by AI bore only a faint10relation to its subsequent returns, and a risk-adjusted long--short strategy earned no reliable11alpha. After the launch of ChatGPT in late 2022, the same disclosure became strongly priced: the12high-minus-low AI-disclosure return spread jumped from roughly $+2$ to roughly $+15$ percentage points13per year, and the long--short Sharpe ratio rose from 0.15 to 1.33.1415The break is robust to signal definition, horizon, dimension reduction, and the exclusion of mega-cap16winners; it is specific to AI-related language (ESG and climate disclosure, tested identically, were17\emph{penalized} in the same window); it is corroborated by a market-model event study in which18high-AI firms earned $+10$ points of additional abnormal return in the year after ChatGPT; and it19operates as a cross-sectional re-rating of persistent firm- and sector-level exposure rather than a20reward for year-to-year wording, fading under firm fixed effects.2122The economic message is that markets price the \emph{language} of AI---but only once AI commands their23attention. The methodological message is that embedding-based measurement of disclosure content opens24a productive avenue for asset pricing, capturing economically meaningful, multi-faceted exposures at25scale and revealing pricing dynamics that keyword counts and firm-differenced designs would miss. As26generative AI continues to reshape the corporate landscape, the semantic footprint it leaves in27mandatory disclosure offers a rich, and now demonstrably value-relevant, signal. Promising next steps28include extending the panel beyond the S\&P~500 to mid- and small-cap firms where attention frictions29are larger, incorporating earnings-call and patent text as complementary AI-exposure measures, and30tracking whether the post-2022 premium persists, decays, or reverses as the AI narrative matures.31