Google disables an AI feature in Google Earth that lets users alter satellite imagery using text prompts. The capability, introduced on Thursday, allowed people to generate or modify images of real-world locations based on natural-language instructions. Reports from multiple outlets say the feature effectively enabled users to create “deepfakes” of the real world by adding or changing elements in satellite views. Within roughly 24 hours, Google removes the function after users used it to depict disaster-related and otherwise misleading scenes, including images that portrayed refugees near the Mexican border and other fabricated disaster scenarios. The company’s action is described as a shutdown of the specific image-editing tool rather than a broader change to Google Earth. Google does not keep the feature available long enough for sustained public use, and the discontinuation follows immediate user-generated misuse concerns. The shutdown underscores how quickly the tool was pulled after users demonstrated the potential for creating misleading satellite-style imagery using AI text prompts.