Google Adds AI Image Tool to Earth, Raising Misinformation Risks on Trusted Maps
Updated
Updated · BBC.com · Jul 31
Google Adds AI Image Tool to Earth, Raising Misinformation Risks on Trusted Maps
3 articles · Updated · BBC.com · Jul 31
Summary
Thursday's rollout lets users place AI-generated scenes onto Google Earth's real satellite, aerial and 3D imagery, then download and share the composites elsewhere.
BBC Verify created false images including a collapsed Eiffel Tower, Russian tanks in Kyiv and a sinkhole swallowing the Great Pyramid, while experts warned the real coordinates lend fake scenes extra credibility.
Google says the images carry invisible watermarks and harmful-topic safeguards, but BBC Verify found prompt tweaks could bypass some restrictions and even trick Gemini into labeling fake Earth images as real.
Researchers said fabricated imagery in conflict zones or hard-to-reach areas could be especially destabilizing because Google Earth is widely treated as a trustworthy source and access to genuine satellite data is already tightening.
If the world's most trusted digital map can now generate fake war zones, how will we ever verify reality during a crisis?
Why would a tech giant risk shattering decades of geospatial trust just to let users paint imaginary scenes over real-world coordinates?
From Pompeii to Propaganda: Google Earth’s 2026 AI Image Generation and the Battle for Geospatial Truth
Overview
In July 2026, Google integrated its advanced Nano Banana 2 image generation model into Google Earth, letting users worldwide create high-precision, AI-generated images anchored to real-world locations. This breakthrough blends generative AI with Google’s vast satellite and terrain data, making it easy for anyone to overlay realistic scenarios onto authentic maps. While this unlocks powerful new tools for planners, educators, and creative users, it also erases the barrier to producing convincing fake satellite imagery. As these fakes inherit the credibility of trusted maps, public trust in satellite data erodes, and technical safeguards like watermarks and metadata often fail to prevent the spread of misinformation.