Nano Banana image generation comes to Google Earth

Google Earth has introduced a generative image feature called Nano Banana, which lets users produce AI-created depictions of locations as they may have looked at earlier points in time. The feature is framed around "visualizing the past," giving everyday users a way to engage with historical context through generated imagery rather than relying solely on archival photographs or satellite data.

The integration is notable because it brings image generation into a product used primarily for navigation, exploration, and geographic reference - not creative work. By anchoring the feature to real places and historical framing, Google appears to be positioning generative AI as a practical tool for curiosity and education rather than open-ended content creation.

Nano Banana itself is a compact image generation model, with the "Nano" designation suggesting it is optimized for efficiency and on-device or lightweight deployment. This fits a broader trend among AI developers toward smaller, faster models that can be embedded into existing products without requiring heavy infrastructure. Details on the underlying architecture or training data have not been fully disclosed.

For Google Earth specifically, the addition continues a pattern of the platform layering new data visualization capabilities over time - from 3D terrain and Street View to timelapse satellite imagery. AI-generated historical visuals add a speculative dimension to that history, which will likely prompt questions about accuracy and how clearly the product distinguishes generated images from documentary ones. How Google handles that transparency will be worth watching as the feature rolls out more broadly.