Google Earth AI Image Generation Paused: Fictional Scenes Overlaid on Satellite Images Risk Spreading Misinformation

Google has announced the suspension of its AI image generation feature on Google Earth less than 48 hours after launch, citing the need to strengthen safety protections. Experts warn that this tool allows fictional scenes to be superimposed onto real satellite, aerial, or 3D maps, making it easy to use for spreading false information.

According to a report by the UK's BBC, the new feature integrated Google's image-generating AI model, Nano Banana 2, into Google Earth. Users could input text prompts to generate fictional scenes directly onto real satellite imagery, aerial photos, or 3D maps, then download and share them on other platforms.

The BBC’s fact-checking team tested the feature and successfully generated fake images, including the Eiffel Tower collapsing, a massive sinkhole swallowing Egypt’s pyramids, and Russian tanks appearing in Kyiv.

Experts point out that when fictional content is combined with real geographic coordinates and Google Earth’s base map, it becomes easy to mislead people into believing the images originate from reliable satellite data. During times of war or disaster, when information is chaotic, such images could particularly mislead journalists and the public.

Google stated that AI-generated images contain invisible watermarks, which users can verify using Gemini or Google Lens. However, BBC testing found that some verification mechanisms could be bypassed, and external detection tools sometimes failed to identify fake images.