Google LLC Rolls Back Google Earth AI Feature Over Misinformation Risks

Google LLC rolled back an AI-powered image generation feature in Google Earth on July 31, 2026, less than 48 hours after its launch. The tool, which integrated the Nano Banana 2 image generator, allowed users to overlay synthetic imagery—such as Russian tanks in Kyiv or a collapsed Eiffel Tower—onto real satellite, aerial, and 3D maps using text prompts.

The company paused the feature following reports from BBC Verify and demonstrations by open-source researchers. Expert Henk van Ess showed that the tool's safety filters could be bypassed to create deceptive images of a nuclear plant in Iran, refugees at the Mexican border, and a bomb crater near a hospital in Gaza. Other tests generated a flooded U.S. Capitol and a crash in Amsterdam.

While Google LLC initially defended the tool by citing the use of SynthID invisible watermarks and safety guidelines to block harmful topics, critics argued these protections were insufficient. Experts from Bellingcat and the Washington Post warned that the ability to easily fake satellite imagery could undermine the reliability of geospatial data used to verify global atrocities and allow governments to dismiss authentic evidence as fake. Google LLC acknowledged that users were sharing imagery that violated company policies and stated the rollback was necessary while the company works to implement stronger guardrails.