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, precipitation, wind, and atmospheric pressure up to 10 days in advance in a matter of seconds, a process that typically takes hours for traditional supercomputer-powered models. Initial testing shows the AI system achieves a 15-20% improvement in predicting key variables like temperature and wind speed at certain altitudes. While promising, meteorologists note that AI models must still prove their reliability in predicting extreme weather events, which are less common in training data. The technology represents a significant shift toward data-driven weather prediction that could make forecasts more accessible and faster to produce. Read the full article at https://technologyreview.com/ai-weather-model-forecast-accuracy.