Google Removes AI Image-Editing Feature from Earth App After Misuse

Google disabled an AI-powered image-editing feature in its Earth application less than 24 hours after it was introduced. The tool, which used the Gemini Nano Banana 2 model to alter satellite views based on text prompts, was quickly employed to produce fabricated scenes such as fictitious refugee camps, a non-existent flood, and a simulated attack on a nuclear facility.

"We've seen geospatial professionals using this feature for a range of useful purposes, however we've also seen people sharing screenshots of generated imagery that appear to violate our policies," a Google spokesperson said in a statement posted on X. The company said the generated images carried a watermark and were never displayed in the public Google Earth experience, and it is pausing the feature while it develops stronger safeguards.