Chicago Booth Study: GPT-4 Outperforms Financial Analysts in Earnings Predictions

May 24, 2024, 08:18 PM

A recent study by Chicago Booth has demonstrated that GPT-4, a large language model, can outperform financial analysts in predicting the sign of earnings changes when fed with financial statements. The research, which has been trending on HackerNews, shows that GPT-4 generates useful insights from its analysis of trends and financial ratios, performing on par with narrowly specialized models. The study suggests a significant disruption in the financial analysis industry, highlighting the potential of GPT-4 to handle tasks traditionally managed by human analysts. Notably, this is achieved despite only feeding GPT-4 with numbers, and the study has garnered attention from BBVAResearch and ECB_Research.