This summer, 38 million farmers in India received AI-generated forecasts indicating the onset of the monsoon season. These predictions utilized NeuralGCM, a model developed by Google Research that integrates physics-based techniques with machine learning for precise forecasting. The University of Chicago implemented NeuralGCM alongside various advanced models to anticipate the monsoon up to a month ahead, successfully predicting an unusual dry period in 2025. In partnership with the Indian Ministry of Agriculture, these forecasts were communicated through SMS, aiding farmers in decisions regarding sowing times, seed purchases, or crop changes. Reliable early forecasts like these can greatly enhance farmers’ earnings and bolster climate resilience in rural areas.
