- GPT-4 predicts company earnings more accurately than human analysts.
- The AI matches advanced machine learning models using only raw financial data.
- GPT-4 excels in complex situations, particularly with small or volatile companies.
Recent research reveals that OpenAI’s GPT-4 can predict changes in company earnings more accurately than human analysts. The study compared GPT-4’s predictions with those of human analysts and advanced machine learning models, using only raw financial data. Surprisingly, GPT-4 not only matched but also outperformed human analysts in accuracy and prediction relevance.
The researchers tested GPT-4 by providing it with standardized financial statements without company names or dates to eliminate prior knowledge influence. Using step-by-step prompts, GPT-4 was guided through the analysis process similar to human reasoning. The AI achieved a prediction accuracy of 60.35%, significantly higher than the 52.71% accuracy of human analysts.
GPT-4 demonstrated particular strength in analyzing small companies and those with volatile earnings, areas where human analysts often struggle. The study suggests that combining insights from both GPT-4 and human analysts can lead to even better predictions, streamlining the financial analysis process and making it faster and more reliable.
