Google added generative image creation to Google Earth, letting a real satellite view be rendered as something else entirely. It's a picture, not a map edit. The feature treats a location's actual satellite, aerial, and 3D imagery as a visual anchor, generates a new image over that base, and leaves the underlying data exactly as it was.
To try it, you zoom into a location in Google Earth on the web and select the create image option, then describe what you want to see. Google's own examples include rendering the Pompeii ruins as a bustling first-century street, turning an empty lot in Tokyo into a retail district, and reimagining a familiar office campus as a glowing science-fiction skyline.
According to Google, the model behind the feature is Nano Banana 2, which also draws on Gemini to pull in relevant background details, for a historical scene or an explainer graphic, for example, before rendering the image. Google says the rollout is global and live now for Google Earth's web app; the company's post does not mention the Android or iOS apps, the Google Earth Pro desktop client, or any subscription requirement, so for now this looks like a web-only, free-tier feature.
What's not addressed is how a product whose credibility rests on being an accurate record of the physical world will keep AI-generated scenes clearly distinct from the satellite imagery underneath them, especially once people start sharing the renders outside the app.
