Google has dropped a new AI gizmo that literally puts its Nanobanana image generation tool on the map.
Nano Banana 2, Google's top image-generation model, is now being integrated for use in Google Earth on the web, so that users can deploy prompts that latch onto real-world locations.
It means, according to Google's case study, users like teachers can feasibly render historical reconstructions onto the Google Earth locations, for example, to turn the modern ruins of Pompeii into a hyper-realistic street scene from 78 A.D.
For less academic and more practical uses, architects and urban planners can place proposed developments on vacant parcels, producing photorealistic 3D visualisations such as a reimagined shopping district over an empty lot.
Homeowners can preview bespoke builds by prompting the tool to add a structure, for example a modern lakefront cabin built from local materials, into the actual landscape.
Of course, expect to also see some more creative or speculative (aka 'sloppish') use cases emerge as real-world users get their hands on the tool.
How it works
Google noted that users can zoom to any spot in Google Earth, tap "create image" and type a prompt to generate images that combine the platform's satellite, aerial and 3D imagery with Nano Banana 2.
Users can also ask for on-map infographics; and to deliver those insights, behind the scenes Gemini retrieves relevant historical information while Nano Banana generates a concise graphic.
It added that the new image generation with Nano Banana is now available globally for Google Earth web users.
