Visual Information and AI Divide: Evidence from Corporate Executive Presentations
55 Pages Posted: 26 Jun 2023 Last revised: 18 Jul 2023
Date Written: November 17, 2022
Abstract
This paper constructs and studies a novel data set comprised of corporate executive presentations, which provide unique visual information about firms' products and operations. We explore the value of visual information in presentations and examine how market participants respond to such information. We extract visual features from presentation images with large image models and find not all images have the same implications for firm value. Forward-looking operational information is associated with higher short-term abnormal returns and long-term operational performance while other types of images are not. We also examine whether the rise of alternative big data and AI creates a potential AI divide among market participants. We find AI-equipped institutional investors respond strongly to visual signals, whereas retail investors and traditional institutional investors face marginalization on the playing field in the age of AI and big data.
Keywords: JEL Classification: D83, G12, G14, G23, G30 Machine Learning, Artificial Intelligence, FinTech, Corporate Presentation, Image Analysis, Textual Analysis, Information and Market Efficiency, Institutional Investors, Investment
JEL Classification: D83, G12, G14, G23, G30
Suggested Citation: Suggested Citation