A new study published in Nature demonstrates that artificial intelligence models can produce accurate medium-range weather forecasts faster and with less computational power than traditional physics-based models. Researchers from several institutions developed an AI system trained on decades of historical weather data. The system successfully predicted key atmospheric variables, such as temperature and pressure, up to ten days in advance with accuracy comparable to established numerical weather prediction methods. The findings suggest AI could significantly reduce the cost and energy required for high-resolution forecasting, though experts note further testing is needed for extreme weather events. For the full details, read the complete article at https://sciencedaily.com/releases/2024/07/240715140810.htm.