Google Earth AI maps outbreak risks with geospatial agents

What is changing?

Google introduced Earth AI-based geospatial reasoning and disease-forecasting prototypes on October 6. The system combines satellite imagery, environmental signals, mobility, and population models with local health data. The WHO AFRO team used the agent during the Ebola response in the Democratic Republic of the Congo to map mining corridors with high exposure risk and human mobility. According to Google, the team identified 48 settlements and more than 45,500 people at risk within minutes, a process that could otherwise have taken weeks.

The same approach was evaluated for weekly forecasts of Ebola spreading into new zones, dengue surges in Mexico, and cholera hotspots in the DRC. Population Dynamics Foundation Model embeddings are available in preview on Google Maps Platform as Population Dynamics Insights, while no-cost access can be requested for selected research use cases. For production teams, the value is letting specialists conduct geospatial analysis in natural language; results still need validation against field data and local public-health expertise.