Google Earth Pulls AI Image Generator After Users Created Misleading Images

Google Earth has temporarily disabled its Nano Banana AI image generator after users exploited it to create misleading satellite imagery, including toppled landmarks, fake nuclear plants, and fabricated refugee scenes. The feature, introduced just days earlier, was pulled following reports from digital misinformation expert Henk van Ess, who demonstrated on his X page how easily the tool could generate convincing but false images, such as a fake hospital and bomb crater in Gaza. Google acknowledged the issue, stating that the feature was rolled back to implement stronger guardrails, and emphasized that generated images were watermarked with SynthID and not visible in the main Google Earth experience. However, tests by the BBC revealed that the watermark could be circumvented, and an X community note highlighted that SynthID only covers sections of AI-generated media, allowing for verification bypass. Google plans to reintroduce Nano Banana once safeguards are improved. The incident raises concerns about the trustworthiness of Google Earth as a reliable source of satellite imagery, especially given the potential for misinformation in geopolitically sensitive contexts.