Google Earth Rolls Back AI Image Tool Over Fake Satellite Photo Concerns

Google temporarily rolled back an artificial intelligence image editing feature on Google Earth just 24 hours after launch, according to statements from the company and reports by major news outlets on October 30. The feature, powered by the “Nano Banana 2” generative model, allowed users to generate and modify specific geographic regions using satellite, aerial, and 3D map data. Google stated the tool was designed to help visualize historical scenes, real estate developments, and construction projects. However, open-source intelligence (OSINT) researchers and disinformation experts quickly raised concerns that the tool could be exploited to generate deceptive imagery over real terrain.

Disinformation Risks in Satellite and Map Data

Following the feature’s release, independent tests conducted by Agence France-Presse (AFP) demonstrated that the tool could fabricate convincing fake satellite imagery. According to AFP testing, the model successfully generated altered visuals depicting simulated explosions and damage resembling a blast in Paris, a nuclear facility in Iran, a bomb crater in Russia, and an Islamic State training camp in Syria. Because Google Earth serves as a primary baseline reference for journalists, researchers, and investigators tracking conflict zones, natural disasters, and restricted military sites, experts warned that doctored visual data could severely undermine open-source verification efforts.

Google’s Rollback and Safety Response

In response to the mounting criticism, Google product manager Brian Horowitz announced that the company was reverting the update. According to Horowitz’s public notice issued within 24 hours of the rollout, instances emerged where generated images appeared to violate platform policies. Google confirmed that the feature will remain offline until the company builds stronger safety guards to prevent the creation and sharing of misleading geographic content.

Frequently Asked Questions

Why did Google remove the AI image feature from Google Earth?

According to Google product manager Brian Horowitz, the company pulled the feature after noticing instances where generated images violated safety policies by creating misleading or inappropriate visual content over real locations.

What model powered the Google Earth editing feature?

The feature utilized Google’s “Nano Banana 2” generative AI model to let users edit and create imagery based on underlying 3D, aerial, and satellite map data.

Who raised concerns about the tool?

Open-source intelligence (OSINT) experts and disinformation researchers flagged the feature, pointing out that it could be used to fabricate convincing visuals of explosions, military sites, and disaster zones.