Economists run a secret prediction game each year. When ChatGPT took part, here's what happened. (www.businessinsider.com)

In a unique experiment, OpenAI's ChatGPT took part in the annual forecasting contest held by economist David Seif, competing against 160 humans—including economists, hedge fund investors, and tech executives. Participants predicted around 30 events spanning politics, business, and pop culture, with scoring reflecting the accuracy of their likelihood estimates. Notably, ChatGPT quickly absorbed the complex rules and generated predictions within minutes, highlighting its potential for rapid analysis in comparison to the extensive time other contestants required. The results were mixed for ChatGPT, finishing 80th and scoring poorly against a benchmark for predictive accuracy set by Seif. While the AI excelled in areas with ample data—correctly forecasting certain events—it struggled with less predictable scenarios, often missing crucial real-time information that affected outcomes. Despite being outperformed by many contestants, some experts suggest ChatGPT's speed and evolving capabilities may enhance its future predictive performance. The contest illustrates the ongoing intersection of AI tools and human intuition, raising questions about the potential of leveraging AI for large-scale forecasting tasks in various industries.