A new study published in Nature demonstrates that artificial intelligence models can produce accurate medium-range weather forecasts faster and more efficiently than traditional physics-based systems. Researchers developed an AI model trained on decades of historical weather data, which successfully predicted global atmospheric conditions up to ten days in advance. The AI system matched the accuracy of established numerical weather prediction models while completing forecasts in minutes rather than hours, requiring significantly less computational power. The findings suggest AI could revolutionize operational forecasting by providing rapid, high-fidelity predictions, though experts caution that further testing is needed for extreme weather events. Read the full article: https://sciencedaily.com/releases/2024/07/240715140810.htm