A European Central Bank research team analyzes over 22,000 S&P 500 earnings call transcripts using GPT-4o–driven sentiment analysis to measure firms’ generative AI exposure. They categorize mentions into Opportunity, Adoption, and Risk themes, and find that early, sustained AI discussions correlate with up to a 0.62% increase in quarterly stock returns, demonstrating that investor sentiment around AI discourse can boost valuation beyond earnings expectations.

Key points

• Analyzed 22,000 S&P 500 earnings call transcripts (2014–2024) using GPT-4o for sentiment classification into Opportunity, Adoption, and Risk themes.
• Found each one percentage point increase in GenAI discourse associates with a 0.62% lift in quarterly excess returns; Early Exposed firms gain an extra 0.26 interaction boost.
• Documented a sharp post-ChatGPT surge in AI mentions, particularly in IT, and linked thematic shifts from Opportunity to Adoption to stock performance.

Why it matters: This research shows that proactive AI communication is a powerful market signal, offering firms strategic advantage and informing investors' valuation decisions.