Loading...

The Impact of Innovation and AI Semantic Orientation on Firms’ Market-Based Performance

Sharifi, Mahoor | 2025

23 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 58862 (44)
  4. University: Sharif University of Technology
  5. Department: Management and Economics
  6. Advisor(s): Miremadi, Iman
  7. Abstract:
  8. As artificial intelligence has increasingly emerged as a general-purpose technology, the importance of assessing firms’ strategic engagement with AI has grown for both investors and policymakers. From the perspective of signaling theory, corporate disclosures can convey information about firms’ strategic orientation, organizational capabilities, and future growth prospects. However, the existing empirical literature has largely measured firms’ AI involvement using keyword-count approaches or patent-based indicators, which often fail to distinguish semantic content and strategic framing from observable technological activity. Consequently, an important gap remains as to whether firms’ semantic orientation toward AI in disclosed texts is associated with market-based valuation, independent of the intensity of their technological activity. This study addresses this gap by examining the relationship between AI semantic orientation and firm value within the frameworks of signaling theory and the general-purpose technology literature. Using a firm-year panel of publicly listed U.S. companies, we construct an AI semantic orientation index by computing cosine similarity between the embedded vector representation of the first section of annual reports and an AI reference vector. The dependent variable is a winsorized Tobin’s Q, employed as a forward-looking measure of market valuation. We estimate fixed-effects models with firm and year effects and cluster-robust standard errors at the firm level, while controlling for technological activity through patent intensity scaled by total assets. The results show that higher AI semantic orientation is associated with a statistically significant increase in Tobin’s Q. Patent intensity relative to assets also exhibits a positive and significant effect on market-based valuation. Robustness checks further confirm the stability of the findings. Overall, the evidence indicates that capital markets respond not only to observable technological activity but also to the semantic content of corporate disclosures and firms’ strategic positioning toward emerging technologies. By introducing a scalable semantic measure of technological orientation derived from annual-report texts, this study contributes to the innovation and firm-valuation literatures and highlights the potential of financial text analysis for measuring strategic constructs at the firm level
  9. Keywords:
  10. Artificial Intelligence ; Signaling ; Patent ; Tobin’s Q Ratio ; Semantic Embeddings ; Financial Text Analysis

 Digital Object List

 Bookmark

No TOC