Google Pulls Nano Banana AI Tool From Earth Over Deepfake Satellite Imagery Risks
In brief
- Nano Banana, an AI image generation tool, launched on Google Earth July 30, generating satellite imagery from text prompts.
- The tool produced convincing fake satellite scenes—blast craters, flooded landmarks, refugee camps—without content safeguards.
- Researchers demonstrated it could create images of sensitive locations like nuclear plants, raising misinformation and verification concerns.
- Google removed the feature by July 31, citing policy violations and plans to reinstate only with stronger protections.
The Tool and Its Risks
Google launched Nano Banana on July 30 within Google Earth's web version. Users could zoom to any location and click to generate fake satellite scenes from a text prompt. The tool produced convincing imagery of events that never occurred—blast craters in Los Angeles, flooded landmarks, and other geospatial fabrications.
Open-source researcher Henk van Ess tested the tool with prompts about refugees at the Mexican border and a nuclear plant in Iran. None were refused. The speed and ease of generation alarmed journalists and investigators who depend on satellite imagery as a trusted anchor for breaking news.
Why This Matters
Satellite imagery has historically been difficult to fake, making it one of the most reliable tools for verifying atrocities and geopolitical events. Google's tool threatened that trust. Bellingcat researcher Jake Godin warned that one-click generation would streamline creation of fakes and accelerate their spread, adding that governments could now dismiss authentic images as fabricated.
Google included a SynthID watermark on generated images to flag them as AI-generated in tools like Gemini. Critics called that insufficient, arguing few people stop to verify images before sharing them.
What's Next
Google removed the Nano Banana tool by July 31, one day after launch. The company said it would reinstate image generation in Google Earth only after implementing tighter protections, though it gave no timeline. The rollback signals the tension between AI capability and the erosion of verification tools in an era of synthetic media.
Frequently asked questions
Why is fake satellite imagery dangerous?
Satellite imagery has long been one of the most trusted tools for verifying breaking news and atrocities precisely because it was difficult to fake. Fake satellite imagery could flood the internet with convincing false scenes and erode trust in authentic geospatial evidence, allowing governments and bad actors to dismiss real images as fabricated.
What safeguards did Google include on Nano Banana?
Google included a SynthID watermark on generated images to flag them as AI-generated in tools like Gemini. However, critics argued this was insufficient because few people verify images before sharing them online.
When will Google bring back image generation in Earth?
Google said it would reinstate image generation in Google Earth only after implementing tighter protections, but it gave no specified timeline for when that would happen.
