Bloated Disclosures: Can ChatGPT Help Investors Process Information?

Generative AI tools such as ChatGPT are expected to disrupt numerous industries and could fundamentally alter the way economic agents process information. We probe the economic usefulness of these tools in extracting information from complex corporate disclosures using the stock market as a laboratory. We use the GPT language model to summarize textual information disclosed by companies in their annual reports (MD&A) and during conference calls. Unconstrained summaries are dramatically shorter compared to the original disclosures, whereas their information content is amplified. When the originals have a positive (negative) sentiment, the summary becomes more positive (negative). More importantly, the summaries’ are more effective in explaining stock market reactions to the disclosed information. Motivated by these findings, we propose a novel measure of disclosure “bloat.” We show that bloated disclosure is associated with adverse capital market consequences, such as lower price efficiency and higher information asymmetry. Finally, we show that the model is effective at targeted summaries that distinguish between financial and non-financial (ESG) performance.