Google Earth's AI 'Nano Banana' Banished After Just One Day Due to Fake Images

Google has abruptly pulled its new AI image-generation tool, 'Nano Banana,' from Google Earth, just one day after its highly anticipated launch. The company, under Alphabet, confirmed the swift removal on July 31, 2026, citing widespread user-generated content that violated its policies and fueled serious misinformation concerns. The tool, pitched to help users visualize history, instead quickly became a vehicle for creating convincing, fake scenarios layered over real-world satellite imagery. The Nano Banana 2 AI model allowed users to type simple text prompts and generate photorealistic scenes over any location on Google Earth web version. However, instead of historical visualizations, users quickly created alarming fabricated satellite images, including fictional blast craters in Los Angeles, a flooded U.S. Capitol, and a fire on Iran's Kharg Island. This immediate misuse drew sharp criticism from open-source intelligence experts like Henk van Ess and Bellingcat's Jake Godin, who warned that the tool could severely undermine the trust placed in Google Earth as a reliable source for authentic geospatial information. Google stated it is rolling back the feature to implement 'stronger guardrails,' acknowledging the challenge of balancing AI innovation with preventing misuse. While Google initially claimed all Nano Banana images were watermarked with SynthID for detection, community notes on social media platforms quickly highlighted that this wasn't always effective, limiting full verification. This incident underscores the ongoing struggle for tech giants to deploy powerful AI tools responsibly, especially when they touch on the critical area of digital truth and visual evidence.