Artificial intelligence (AI) is enabling a revolution in weather forecasting, creating the potential to empower farmers with information for agricultural decisions. Multiple technology firms and research institutions have developed global AI weather models and made them open access. It is relatively straightforward to check how well these models predict particular weather variables that are important for agricultural decision-making in particular times and places. The best models for any particular purpose can then be disseminated to farmers, either through mass media or digitally.

Multiple randomized trials find farmers adjust their behavior in response to accurate weather forecasts. Farmers share information with each other, generating benefits of up to $141 USD per farmer.

India digitally distributed AI-based weather forecasts to 38 million farmers in 2025, with dissemination costs less than $0.005 USD per farmer. Forecast messages were developed through focus groups and A/B tests to boost farmer comprehension. The forecast successfully predicted an unusual monsoon year, including a pause in the northward progression of the monsoon. Farmers had not received any other forecasts of the pause in monsoon progression.

Multilateral development banks have committed funds to support other countries who wish to scale weather forecast dissemination. A variety of organizations have provided training in AI weather forecasting, benchmarking of open-access weather models to test how well they predict agriculturally-relevant phenomena in particular contexts, and technical assistance in developing and testing digital messages for conveying forecast information.

More on India’s 2025 dissemination of AI-based forecasts: Press Release from the Ministry of Agriculture & Farmers’ Welfare | The Statesman | The Economist | Wall Street Journal | Human-Centered Forecast Initiative