Google Pulls Earth AI Feature After Taboo Image Generation

Google has quietly discontinued its generative AI feature for Google Earth, a tool intended to overlay AI-generated scenes onto real-world imagery, less than a day after its rollout. The decision came after users began generating taboo, misleading, and disinformation-focused imagery, forcing the company to pull the feature. A Google spokesperson stated to Fortune that the company understands users rely on Google Earth for accurate views of the world and clarified that the generated images did not appear in the main Google Earth experience for others and were watermarked as AI-generated. However, it became apparent that users were exploiting the system to create false images intended for disinformation campaigns. Examples cited include generated images of a nuclear plant in Iran and refugees near the Mexico-US border. The feature allowed users to overlay AI-generated scenes onto real satellite, aerial, and 3D imagery. While Google envisioned legitimate applications such as visualizing planning concepts or historical reconstructions, users rapidly began generating fabricated disasters, false military installations, and conflict scenes linked to actual geographical locations.

Google's Nano Banana technology, integrated into its Gemini chatbot, enables users to generate AI images through simple text prompts. This technology differs from earlier AI image generators by allowing users to edit images, similar to Photoshop, where the AI could add, remove, or alter objects while maintaining realistic features. Google implemented safeguards for the Nano Banana software integration with Google Earth. According to a Google spokesperson speaking with Fortune, images generated by the system were kept separate from authentic Google Earth imagery and were embedded with watermarks to signify their AI origin. The company also utilizes "SynthID," a cryptographic watermarking technology designed to identify AI-generated content. The rapid discontinuation highlights the challenges companies face in balancing the creative potential of generative AI with the risks of misuse, particularly when applied to sensitive or real-world contexts. The incident underscores the ongoing debate and development surrounding AI safety, content moderation, and the ethical deployment of powerful generative technologies. The company's swift action suggests a low tolerance for the misuse of its platforms, especially concerning the potential for spreading misinformation or creating harmful content. The feature's brief existence and immediate withdrawal serve as a case study in the unpredictable outcomes of integrating advanced AI into widely accessible consumer products.