By: Yael Maguire, VP & GM, Google Maps Platform & Google Earth
Today, we’re introducing new generative AI capabilities for Google Maps Platform and Google Earth to provide developers and urban planners with fresh, real-world information and to help solve geospatial problems.
Helping developers build generative AI applications with fresh information about places
More than 10 million websites and apps use Google Maps Platform to offer helpful products, from delivery apps to real estate websites. As more developers build generative AI applications, we often hear that they want an easy way to show up to date information in their products.Today we’re introducing grounding with Google Maps, a new capability that allows developers to use our rich place data to build generative AI-powered experiences. Grounding with Google Maps helps make answers generated by LLMs more factual and fresh, quickly reflecting updates from the 250 million places in Maps.
Imagine you’re a developer for a property rentals website. Through your website’s chatbot, a customer could ask “What fitness options are near this apartment?” Using grounding with Google Maps, you’ll be able to show the latest information about the area, including reviews and business hours for nearby gyms, in addition to how early or late the place is open — making it easy for the customer to make an informed decision. Grounding with Google Maps launches as an experimental feature today.
Launching soon: AI summaries of places in Rivian vehicles
At Google I/O this year, we announced Gemini capabilities in our Places API, which lets developers show generative AI summaries of places. Next month, Rivian will start using this feature to show summaries of restaurants, shops and supermarkets on their infotainment screens — so drivers can quickly and easily learn more about a place.
Bringing generative AI capabilities to Google Earth
We’re also bringing Gemini to Google Earth, which will allow urban planners to access deeper, city-level insights and significantly reduce the time spent analyzing data from days to minutes.
Imagine a transportation planner wants to install new electric vehicle (EV) chargers in their city. All they need to do is ask, “Can you map the five zip codes with the fewest EV chargers relative to their geographic area size?” Google Earth uses multi-step reasoning to determine the five zip codes with the fewest EV charging stations, and then generates a helpful custom visualization. You can also take this query a step further and ask if there are hotels and shopping centers that don’t have EV chargers within a certain distance.
We built this capability in partnership with Google Research and X, the moonshot factory. Next month, we’ll start piloting Gemini capabilities in Google Earth — sign up for consideration to be a trusted tester.
We’re excited to see what developers and urban planners can do with these latest technologies.
