A new artificial intelligence weather forecasting model developed by researchers has demonstrated superior accuracy compared to conventional physics-based systems in global weather prediction. The AI model, trained on decades of historical weather data, can generate high-resolution forecasts for temperature, pressure, wind, and humidity up to 10 days in advance in a matter of seconds, a process that typically takes hours for traditional supercomputer-powered models. Initial tests show the AI system matches or exceeds the accuracy of leading global forecasting models, particularly in predicting extreme weather events. While promising, meteorologists note that AI models currently rely on traditional systems for initial data and may struggle with unprecedented weather patterns not found in their training data. The development signals a potential shift toward hybrid forecasting systems that leverage the speed of AI and the physical understanding of conventional models. Read the full article at https://technologyreview.com/ai-weather-model-forecast-accuracy.