Google Earth AI tool that let users fake satellite imagery pulled after one day amid misinformation backlash

Google launched and then retracted an AI feature in Google Earth on July 31, 2026, after researchers demonstrated that the tool could be used to generate convincing fake satellite imagery. The feature, which integrated Google's Nano Banana 2 image generator, allowed users to create AI-modified versions of real locations with text prompts.

The feature was live for less than 24 hours before Google disabled it. During that window, independent investigators created images showing a giant golden statue of President Donald Trump looming over the White House, refugees near the Mexican border, a nuclear plant in Iran, and a fatal crash on a street in Amsterdam.

Misinformation fears sparked immediate backlash

Researchers who rely on Google Earth to verify images and videos of locations were alarmed. Eliot Higgins, founder of Bellingcat, posted on Bluesky: "Google Earth, a tool often used as a source of satellite imagery to verify photos and videos, has added a feature that allows you to alter satellite imagery with AI, for reasons. No way this could be abused." Henk van Ess, an independent investigator, wrote in a blog post: "What on earth is Google doing?" after generating multiple false scenarios.

• Higgins created an AI-modified image of a giant golden Trump statue at the White House using the tool.
• Van Ess generated images of refugees near the Mexican border, a nuclear plant in Iran, and a fatal crash in Amsterdam.
• BBC Verify was able to create images of a collapsed Eiffel Tower, a sinkhole swallowing the Great Pyramid of Giza, and Russian tanks in Kyiv.
• Google had promoted the feature for use cases such as depicting completed real estate projects or transforming Pompeii into a hyper-realistic view of 78 CE.

Watermark safeguards questioned

Google initially defended the feature by pointing to its SynthID digital watermark system, which automatically embeds invisible watermarks in AI-generated images. The company stated that users could check images via the Gemini app or Google Lens. However, critics noted limitations. Ars Technica previously found that SynthID limits image checks to about 10 per day. BBC testing showed that some third-party detection tools could be fooled into labeling manipulated Google Earth images as real. Van Ess questioned why Google's policy on preventing harmful content creation did not catch the obvious misuse cases before launch.

Google announced the rollback on X, stating it was "continually updating our protections." The company did not provide a timeline for when or if the feature might return. The incident underscores the tension between AI product innovation and the risk of enabling misinformation at scale, especially when the tool is built into a platform that is itself a trusted source for verification.