Google Earth's AI image generation feature, powered by Nano Banana, allowed users to create reality-warping images from satellite and 3D imagery using text prompts. The feature was quickly rolled back after Digital Digging's Henk van Ess demonstrated misleading outputs, including fabricated images of refugees near the Mexican border and a bomb crater near a Gaza hospital. Google defended the tool by citing SynthID watermarks and suggested users verify AI-generated content through Gemini or Google Lens.
Background
Google Earth has long been a trusted source for satellite and aerial imagery, making AI-altered versions particularly concerning for misinformation. The rise of generative AI tools integrated into mapping platforms raises significant questions about verification and trust in visual data.
- Source
- The Verge
- Published
- Aug 1, 2026 at 01:05 AM
- Score
- 6.0 / 10