Google Suspends Newly Launched Earth AI Feature Amidst Disinformation Concerns
Google has abruptly suspended its newly launched AI image-generation feature inside Google Earth, less than 24 hours after a global rollout. This decision came after users rapidly exploited the feature to create fabricated and alarming imagery, raising serious concerns about disinformation and geopolitical manipulation.
The feature, powered by Nano Banana 2, Google’s Gemini-based image generation model, was added to Google Earth on the web on July 30, 2026. It allowed users to select any real-world location, input a text prompt, and receive an AI-generated image based on Google’s actual satellite and aerial imagery of that location. The tool was rolled out globally without any waitlist.
However, within hours, users demonstrated how the system could be manipulated to fabricate dramatic scenes, including explosions, protests, and military activity, anchored to genuine geographical coordinates. Critics quickly flagged the tool’s potential to lend Google’s trusted platform credibility to entirely fictional events. One viral example reportedly depicted a fake nuclear plant in Iran. Google confirmed it observed “screenshots of generated imagery that appeared to breach its policies” and pulled the feature while it works on implementing “stronger guardrails.”
Experts raised particular alarm about the tool’s threat to satellite imagery as a tool for humanitarian research. Nathaniel Raymond, head of the Humanitarian Research Lab at Yale University’s School of Public Health, expressed that it would make his team’s already-difficult work documenting atrocities in places like Sudan “a million times harder, for no apparent underlying value.” Google stressed that generated images were watermarked and not visible to other public Google Earth users.