AI Applications in Economic Forecasting

This document is a technical report that examines the role of artificial intelligence (AI), particularly large language models (LLMs), in economic forecasting. It discusses the emerging research on how LLMs can be utilized for predicting economic indicators, with a focus on inflation forecasts. The report outlines findings from the Federal Reserve Bank of St. Louis, which compared AI-generated inflation forecasts to those from professional economists, noting that LLMs generally produced lower mean-squared errors. The report also highlights a study by Turcotte (2024) that applied similar methodologies to various countries, finding that LLMs often matched or outperformed human forecasters. However, it also addresses the limitations of AI in forecasting, such as potential biases from using future information and the tendency of some models to replicate past forecasts. The report concludes that while AI can be a valuable tool for economists, it should not yet replace human forecasters due to its current limitations.