Google removes Earth AI image generator one day after launch amid misinformation concerns

Jul 30, 2026

Google has removed a new AI image generator from Google Earth just one day after its launch due to widespread concerns about misinformation. The feature, which used the Nano Banana 2 model, allowed users to create fabricated images on satellite maps. Researchers quickly demonstrated its potential for abuse by creating realistic fakes of bomb craters, refugee camps, and military movements, prompting Google to pull the tool to implement stronger guardrails.

Nano Banana Google Earth integration

▪On July 30, 2026, Google launched a feature integrating its Nano Banana 2 AI image generator into the web version of Google Earth.
▪The feature allowed users to superimpose AI-generated images onto Google Earth's satellite, aerial, and 3D imagery using text prompts.
▪Google stated it was "rolling back this feature in Google Earth while we work on implementing stronger guardrails."
▪Google removed the AI image generation feature on July 31, 2026, less than 48 hours after its launch.

Architectural rendering limitations

▪The tool's "Create Image" button was unavailable in the higher-fidelity Street View, limiting its usefulness for detailed work.
▪In an architectural rendering test, an attempt to add a porch to a house resulted in the AI picking the wrong house.

Historical visualization experiments

▪The tool was also used for entertainment, creating images of zombies and evil clowns in Philadelphia and a steampunk version of the city.
▪Experiments to visualize historical scenes, such as Independence Hall in 1776, produced images with an appropriate look but containing anachronistic details.

Location data integration gaps

▪This data integration gap was shown when the AI failed to render the historically accurate Public Ledger building in images of 19th and 20th-century Philadelphia.
▪The feature was described as weak for serious geographic work, treating locations like screenshots rather than using deep map data.

Creative misuse demonstrations

▪Despite watermarks, tests showed it was possible to bypass checks and fool some AI detection tools.
▪To demonstrate the risks, users created fake images of a refugee camp near the Mexican border and a bomb crater by a hospital in Gaza.
▪Other fabricated images included a collapsed Eiffel Tower, Russian tanks in Kyiv, a fake nuclear facility in Iran, and a plane crashing into a Manhattan skyscraper.
▪Google's initial response stated that all generated images contained a SynthID digital watermark to identify them as AI-created.
▪Experts warned the tool could erode trust in satellite imagery, a historically reliable source for journalists and researchers.
▪The removal followed widespread criticism that the tool could be used to create and spread realistic-looking misinformation.