From progress to loss: When AI writes about risk
Companies are increasingly turning to generative AI to draft their reports: shorter texts and simplified language. But does easier reading mean better information?
Listen to your favorite podcast here or read the article here.
Generative AI is transforming the way companies write about risk, shortening lengthy reports and simplifying language that was criticized for years as impenetrable. But does this greater readability also mean better information for investors? In this episode of Risk and Return, we put that assumption to the test.
Together with Giulia Redigolo, Associate Professor in the Department of Economics, Finance and Accounting at Esade, we analyze more than five million sentences drawn from the risk reports of US companies between 2019 and 2024, using the launch of ChatGPT as a natural dividing line. The finding: the more AI is involved in drafting these sections, the less investors seem to learn from them, and that loss of company-specific detail carries a measurable cost in how the market reacts.
Hosted by Professor Omar Rachedi, Risk and Return features conversations with leading academics and practitioners to cut through the jargon and explore how markets, companies, and policies shape our world.