Quick Facts
- Google launched and pulled its AI image-generation tool in Google Earth within 48 hours, citing policy violations.
- Users generated fake satellite imagery including explosion craters in Los Angeles, Russian tanks in Kyiv, and ISIS camps in Syria.
- Researchers and open-source investigators warned the tool could permanently damage public trust in satellite imagery.
Google removed an AI image-generation feature from Google Earth on July 31, 2026, less than 48 hours after launching it globally. The company cited screenshots of generated imagery that violated its policies.
The tool, powered by Google’s Nano Banana 2 image-generation technology, launched July 30 on the desktop web version of Google Earth. Users could zoom into any location, type a prompt, and generate a photorealistic image layered over real satellite, aerial, and 3D mapping data. Google marketed it as a way to visualize history and create real estate plans.
The feature required no waitlist and carried no cost. It was not available on iOS or Android, though the Android app alone has surpassed 500 million downloads.
What made the tool distinct from standard image generators like DALL-E or Midjourney was that every fabricated image inherited real coordinates, lighting, and angles from the underlying map data. A fake explosion crater in Los Angeles sat at accurate GPS coordinates under correct lighting conditions, making it appear far more credible than a typical AI-generated image.
Within hours of launch, users had generated scenes including a submerged U.S. Capitol, Iran’s Kharg Island burning, a collapsed Eiffel Tower, a sinkhole at the Great Pyramid of Giza, and Russian tanks in Kyiv. AFP testing produced explosions in Paris, nuclear sites in Iran, bomb craters in Russia, and ISIS training camps in Syria.
GeoConfirmed, an open-source intelligence volunteer group, posted on X that individuals were already using the tool to fabricate military positions and relabel buildings. “The primary, and perhaps only, practical use of this tool appears to be the creation and dissemination of mis- and disinformation,” the group wrote.
Ross Burley, executive director of the Centre for Information Resilience, called the launch irresponsible. “Trust in satellite imagery has taken decades to build and could be irrevocably damaged overnight,” he told AFP.
Brady Africk, a research analyst at the American Enterprise Institute, said the tool made it “easier to generate convincing fake satellite imagery that can spread quickly online and mislead the public.” He added that such fakes “erode public trust in satellite imagery, and make the jobs of journalists and researchers more difficult.”
Jake Godin, a researcher at Bellingcat, warned that putting AI image generation behind a single click would accelerate the spread of fakes. “Misinformation travels further and faster than any sort of correction,” he said.
Google had built in a SynthID digital watermark on every generated image and argued users could verify authenticity through the Gemini app or Lens in Search. Critics noted most users do not check for watermarks. BBC Verify testing found it was possible to circumvent the watermark checks in some cases, and external AI detection tools also failed to identify certain generated images.
In a statement, Google said: “We know that people uniquely trust Google Earth for a reliable view of the world.” The company added it was rolling back the feature “while we work on implementing stronger guardrails.”
The reversal follows a pattern for Google. In February 2024, the company pulled Gemini’s ability to generate images of people after the tool produced historically inaccurate depictions. The Earth AI rollback happened faster, but both cases point to the same gap: AI features shipped before content moderation systems were stress-tested against real-world abuse.
For technology companies building AI tools on top of trusted data sources, the Google Earth episode sets a clear precedent. The credibility of the underlying platform can become a liability when a generative layer is added without adequate safeguards. Journalists, human rights investigators, and courts have relied on Google Earth as a stable record of physical reality for more than two decades. That institutional trust, once questioned, is difficult to restore.
This article was written by an AI agent. Spotted an error? Send a correction and we will fix it.