Google rolls back AI image generation in Google Earth after misleading satellite imagery

Google has rolled back the Nano Banana image generation feature in Google Earth less than a day after announcing the integration on July 30. The move comes after people used the tool to create and share misleading satellite imagery.

In a statement posted on X, the company said it understands that users uniquely trust Google Earth as a reliable view of the world. While geospatial professionals have used the feature for useful purposes, Google said it also saw screenshots of generated imagery that appear to violate its policies.

The rollback followed a report from 404 Media showing how easy it was to produce deceptive satellite images with the tool. In one example, author Joseph Cox generated an image of a blast crater in Los Angeles. In another, he placed protestors outside Google's Mountain View campus in California.

At first, Google played down the misinformation concerns, arguing that every image created with Nano Banana includes a SynthID digital watermark. The company suggested users could verify an image's authenticity by asking the Gemini app or using Lens in Search. In addition, it said it prevents image creation on harmful topics and continually updates its protections.

Cox later criticized that response, saying the watermark argument misunderstands how misinformation spreads. People often do not stop to verify images they see online, he noted, adding that only a few years ago the internet shared AI-generated baby peacock images as real. On Friday, Google repeated its watermark defense, but also said it would implement stronger guardrails before enabling image generation in Google Earth again.