Google has unveiled a series of artificial intelligence advancements powered by Gemini, integrating the model into both its navigation services and environmental forecasting systems. The company announced that Google Maps is introducing a new conversational feature called “Ask Maps,” allowing users to interact with the app using natural language to find locations, plan trips, and receive contextual recommendations. The AI assistant can respond to complex queries and personalize suggestions based on a user’s search history and saved places.
Alongside the AI assistant, Google introduced Immersive Navigation, a redesigned driving experience featuring 3D views of buildings, terrain, and road infrastructure. The update includes more natural voice guidance, real-time alerts for road disruptions, and improved route comparisons using data from the Google Maps and Waze communities. The feature is rolling out across the United States on iOS and Android, with expansion planned for additional platforms including CarPlay and Android Auto.
In parallel, Google Research has applied Gemini to disaster forecasting by analyzing five million global news articles to identify 2.6 million flood reports, creating a geotagged dataset called Groundsource. The dataset was used to train a flash-flood prediction model that combines global weather forecasts with machine learning to estimate flood risk.
According to Gila Loike, a product manager at Google Research, the approach represents a new application of language models for environmental data analysis. Juliet Rothenberg, a program manager on Google’s Resilience team, indicated that the dataset could help address gaps in regions lacking weather monitoring infrastructure. The forecasting model now highlights flash-flood risks in urban areas across 150 countries through Google’s Flood Hub platform and is being shared with emergency response agencies worldwide.
