GPT-4’s financial analysis superiority heralds AI’s dominance over business analysts

Large language models like GPT-4 have been shown to be able to analyze financial statements with accuracy that is equal to, and in some cases superior to, that of humans, scientists at the University of Chicago have found.

Researchers examined the effectiveness of GPT-4 in analyzing corporate financial statements to predict future earnings trends. “Even without explanations and industry information, a large language model outperforms financial analysts in its ability to predict changes in earnings. The large language model demonstrates a relative advantage over human analysts in situations where analysts struggle. Moreover, we found that the predictive accuracy of the large language model is on par with a highly specialized state-of-the-art machine learning model,” the researchers write.

The research project found that the large language model generated “useful information about a company’s future performance” and that trading strategies based on GPT forecasts “produced higher Sharpe ratios and alpha than strategies based on other models.” Scientists attribute such results to the extensive knowledge base of large language models, as well as the ability to quickly recognize patterns and business concepts – as a result, AI makes “intuitive” decisions even in the presence of incomplete information.

For the foreseeable future, financial analysis will, of course, be the work of people, but as the study showed, large language models can become a useful tool for them that optimizes their work and helps them make informed decisions. The work of financial analysts may change in the future.

Leave a Reply

Your email address will not be published. Required fields are marked *