Google Earth Integrates Nano Banana 2 AI Image Generator

Google has embedded the Nano Banana 2 image-generation model into Google Earth on the web. Users click a “Create Image” button, enter a prompt, and receive an AI-generated picture tied to the map location. The feature is live globally and can produce historical visualizations, infographics, real-estate concept renders, or imaginative scenes. Google’s blog promotes it as a way to “visualize the past” for teachers, developers, and planners.

ForbesEngadgetZDNet

Nano Banana, Google’s multimodal AI model, was previously accessed through Gemini chat sessions, requiring screenshots and a separate interface. Earlier demos showed the model in Earth, but the new integration makes generation a native toolbar function. This follows a broader trend of embedding generative AI into mapping platforms.

ZDNet

• Historical reconstructions – promotional material shows Pompeii rendered as a bustling Roman street and the Statue of Liberty turned into an annotated infographic.

Forbes

• Real-estate planning – an empty lot in Tokyo is reimagined as a “vibrant shopping district,” and a New Jersey house receives a wrap-around porch (the AI mistakenly applied it to a neighboring property).

ZDNet

• Creative experimentation – testers generated whimsical scenes such as Philadelphia overrun by zombies, evil clowns, giant alien mechs, and steampunk-styled Independence Hall.

ZDNet

Reviewers note the AI treats the map view as a static screenshot rather than a spatially aware dataset. It misplaces architectural elements, fails to reproduce accurate historical dimensions, and often inserts incorrect labels or nonsensical captions. In an Independence Hall test, the AI omitted the 19th-century Public Ledger building and produced an inaccurate 1976 façade, indicating limited integration with Earth’s historical data. The tool also lacks a “Create Image” button in Street View, forcing reliance on satellite angles that reduce fidelity.

Engadget

Google markets the feature as an “instant transformation” of modern imagery into “hyper-realistic” historical scenes, emphasizing educational benefits for visual learners. The company claims it can help clients see possibilities in real-estate projects and generate custom infographics for any location, though it has not clarified limits on the infographic capability.

Engadget

If refined, the integration could aid teachers in creating immersive historical visuals, assist developers in early-stage site concepting, and offer a novel way for the public to explore speculative futures. Current limitations—particularly inaccurate historical rendering and poor text handling—confine its utility to novelty and exploratory use cases.

Engadget

Google asserts the infographic feature works on any Earth location, yet demonstrations have only shown high-profile sites such as the Statue of Liberty. Independent testing reveals factual inaccuracies and misplacements, leaving the geographic scope uncertain.

Forbes

Google has indicated future releases may incorporate deeper location awareness, including potential integration with Street View, to improve fidelity and usefulness. The company’s history of iterating on early AI releases suggests the Nano Banana add-on could evolve into a more robust visualization tool.

ZDNet