Google Earth urgently halts AI-generated images: fabricated scenes can be overlaid on satellite images to spread misinformation.

Less than 48 hours after launching the AI image generation feature for Google Earth, Google announced a suspension of the service to enhance security measures. Experts had previously warned that this tool could overlay fictional scenes onto real satellite images, making it susceptible to the spread of misinformation. The BBC reported that this new feature integrates Googles image generator, Nano Banana 2, into Google Earth. Users simply need to input text commands to generate fictional scenes on real satellite images, aerial images, or 3D maps, which they can then download and share on other platforms. During testing by the BBC verification team, they successfully generated false images such as the collapse of the Eiffel Tower, a giant sinkhole swallowing the Egyptian pyramids, and Russian tanks appearing in Kyiv. Experts pointed out that when fictional content is combined with real geographical coordinates and Google Maps basemaps, it can easily mislead people into thinking the images come from reliable satellite data. Such images are particularly likely to mislead journalists and the public during times of information chaos due to wars or disasters. Google stated that AI-generated images contain invisible watermarks, which users can verify through Gemini or Google Lens. However, the BBC testing revealed that some verification mechanisms could be circumvented, and external detection tools sometimes fail to identify fake images.