AI Weather Forecasts Could Help Public Health Plan for Heat and Disease Risks

AI Weather Forecasts Could Help Public Health Plan for Heat and Disease Risks

September 22, 2026

A Climate Week event hosted by the University of Chicago’s Institute for Climate and Sustainable Growth and the Rockefeller Foundation is launching a report on AI-powered weather intelligence for health. The project brings together people working in forecasting, public health, climate science, and social science to consider how better weather information could help protect communities.

Weather affects health through heat, smoke, flooding, infectious disease, and disruptions to care. More detailed forecasts could help hospitals prepare staffing, cities open cooling centers, and public agencies communicate risk earlier. But prediction is only one link in the chain. People need timely warnings, services, and the ability to act on them.

Forecast accuracy is not enough

A forecast can be technically accurate yet fail to help the people most at risk. Public-health use requires information at a useful geographic scale, clear uncertainty, and delivery through channels residents can access. If forecasts depend on data that underrepresent rural areas or vulnerable neighborhoods, the system may perform unevenly.

AI models may improve some forecasting tasks, but they should be compared with established physics-based systems and human forecasters. Evaluation should measure not only average accuracy but also performance during extreme events, when mistakes are most consequential.

Turning information into action

Health agencies need protocols that connect a forecast to a decision. A heat-risk alert could prompt outreach to older adults, but only if agencies know whom to contact and have resources to provide support. A smoke forecast may help clinics prepare for respiratory cases, but staffing and medication supply still need planning.

Local leaders should be involved in designing warning systems. Messages need to be understandable, multilingual, accessible to people with disabilities, and clear about uncertainty. Communities should be able to report when an alert missed local conditions.

Data and trust

Combining weather, health, and location data raises privacy questions. Agencies should use the minimum information necessary and explain how data are stored and shared. People may lose trust if a system appears to monitor individuals without clear public benefit.

The report’s value will depend on recommendations that connect forecasting to real public-health capacity. AI may make weather intelligence more timely, but it cannot replace clinics, emergency response, clean-air shelters, or trusted community organizations. The goal should be a stronger warning-to-action system, not another dashboard without resources behind it.

Sources

• University of Chicago event: Forecasting the Future
• Climate Week NYC event details
• WHO: Climate change and health