General chatbots struggle with planetary data because it is huge and comes in specialised formats. Google’s agent links Gemini to Earth AI’s weather, flood, population and satellite models, writes a step-by-step plan, calls only the tools it needs and shows its plan and data sources. Partners including the WHO’s Africa office, GiveDirectly and the Milken Institute tested it on outbreak investigations, disaster preparation and clean energy screening.
Why it matters
Questions like where a disease will spread or where to build solar farms need many datasets combined correctly. If agents can do that reliably, organisations without specialist mapping teams gain access to planet-scale analysis. Showing the plan and sources matters, because experts need to check the assumptions before acting on the results.
Original source
Putting Earth AI to work: Geospatial agents for our planet’s biggest challenges
Google Research
research.google
Summary by Second Order AI. Full reporting is at the original source.