Google introduced an AI-powered image generation feature within Google Earth on July 30, 2026. By the following day, July 31, the company had removed it entirely.

The feature, named Nano Banana, allowed users to create AI-generated visual overlays on real satellite, aerial, and 3D imagery using text prompts. Shortly after launch, users discovered potentially harmful applications for the tool, prompting Google to disable it.

The rapid reversal reflects a broader tension across the technology industry, where companies including Google, Microsoft, and Meta have been racing to integrate generative AI into consumer products while simultaneously grappling with how to prevent misuse. Google Earth, launched in 2005, has become one of the most widely used geospatial reference tools in the world, relied on by governments, journalists, educators, and everyday users alike.

How Nano Banana Worked

Nano Banana enabled users to generate custom AI images positioned on top of real-world geographic imagery inside Google Earth. Users could enter a text prompt, and the system would produce a corresponding visual overlay layered onto actual locations.

Google positioned the tool as a creative and educational resource. Potential use cases included visualizing historical events at their real-world locations, creating real estate mockups on actual plots of land, and designing futuristic cityscapes overlaid on existing skylines.

The company implemented several guardrails. All generated images carried watermarks identifying them as AI-created content. The images were kept in a sandbox environment separate from the core Google Earth experience, meaning other users browsing the map would not encounter someone else's AI-generated content. The overlays were visible only to the creator.

Swift Backlash Over Misuse Potential

Within hours of the launch, critics identified that Nano Banana could be exploited to fabricate satellite imagery depicting events such as drone strikes, altered national borders, or invented military installations — all rendered on real, identifiable geography.

Satellite imagery holds a unique position of authority. Journalists, researchers, and intelligence analysts have long depended on Google Earth as a credible and neutral source of geospatial information. A tool capable of easily producing convincing forgeries of such data, even with watermarks in place, risks undermining that established trust.

The concern is not theoretical. Over the past several years, AI-generated images have become increasingly difficult to distinguish from authentic photographs, and doctored visual content has already been used in disinformation campaigns. The spread of convincing synthetic media has prompted ongoing industry efforts around content provenance standards, though no universal solution has been adopted.

Google withdrew the feature one day after its debut. While the company highlighted the watermarking and sandboxing protections it had built, it ultimately determined that the reputational risk outweighed any potential engagement benefits.