| dc.description.abstract | The release of ChatGPT constituted a major technology shock to AI, offering new
information about AI advancement and its future potential. This paper studies the
impact of the release on analyst stock recommendations and earnings forecasts for AIexposed
firms, which are argued to have a competitive advantage in capitalizing on
such advancements. We measure AI-exposure with a combination of natural language
processing techniques applied to earnings call transcripts. Using a difference-in-differences
model, we find evidence of recommendations for AI-exposed firms moving towards "sell".
Findings also indicate a positive effect on earnings expectations likely driven by reduced
cost forecasts. However, recommendations indicate that the earnings effect is not strong
enough to justify an upgrade. Thus, we attribute "sell" recommendations to stock prices
moving beyond what analysts consider fundamental values. | en_US |
