Google discontinued a new AI-powered image generation feature within Google Earth following widespread misuse that raised alarms about misinformation. The tool, active for just one day, allowed users to produce realistic satellite-style images based on text prompts, but quickly led to fabricated visuals of bombings, disasters, and protests in real-world locations.
Built on Google’s Nano Banana 2 technology, the feature enabled the creation of photorealistic images overlaying Google Earth’s satellite, aerial, and 3D maps. Users experimented by generating scenes such as nuclear plants in sensitive regions, simulated drone attacks on workplaces, and homeless encampments in politically charged urban areas. Such manipulations motivated critics and investigators to call the tool "irresponsible" and harmful to fact-based research.
Google had positioned the AI function as a visualization aid to explore historical cityscapes, imagine urban projects like community gardens, or plan real estate developments. However, users quickly exploited it to fabricate scenarios that could be mistaken for authentic satellite photos, including war-torn cities and disaster sites. This divergence between intended use and actual practice prompted Google to remove the feature.
Independent researchers and journalists demonstrated how easily the tool generated misleading imagery, including bomb blasts in major cities and protests staged outside corporate buildings. The ability to conjure damaging scenes with realistic detail evoked concerns about the potential spread of false information, especially related to conflict zones. Critics warned that such imagery could be misused to manipulate public perception and undermine credibility.
Though the AI tool showcased innovative possibilities in geospatial visualization, its swift removal highlighted challenges in managing generative AI’s ethical boundaries within widely used platforms like Google Earth. The company did not specify plans to reinstate or modify the feature but acknowledged the need to address user safety and misinformation risks before proceeding further.