Google Pulled Its Google Earth AI Image Tool After One Day of Viral Misuse

What happened

On Thursday, Google released a new AI feature inside Google Earth that allowed anyone to generate fake images and place them on the platform's map interface. Within hours, users began creating and sharing screenshots of fabricated scenes: a fake nuclear plant in Iran, a plane crashing into One World Trade Center in New York, and images of crowds massing at the U.S.–Mexico border.

The backlash was swift. By Friday—just one day after launch—Google announced it was pulling the feature. In a statement posted online, the company said: "We know that people uniquely trust Google Earth for a reliable view of the world. We've seen geospatial professionals using this feature for a range of useful purposes, however we've also seen people sharing screenshots of generated imagery that appear to violate our policies."

Google added: "So we're rolling back this feature in Google Earth while we work on implementing stronger guardrails." The company noted that generated images did not appear in the main Google Earth experience for other users to see and were watermarked as AI-generated—but screenshots shared on social media stripped away that context.

Why it matters

Google Earth is one of the most trusted geographic reference tools in the world, used by journalists, researchers, governments, and everyday people to verify what a place looks like. Introducing AI-generated imagery into that environment—even in a separate layer with watermarks—created an immediate trust problem. Screenshots of fake scenes circulated without their AI labels, making it easy for misleading images to go viral before anyone could verify them.

The episode highlights a recurring pattern in AI product launches: powerful generative tools are released with minimal guardrails, bad actors quickly exploit them, and companies retreat only after public backlash. Google's own statement acknowledged that geospatial professionals found the feature useful, suggesting the underlying technology had legitimate applications—but the company could not control how screenshots of generated content traveled outside its platform.

The bigger question is why Google launched the feature without anticipating the obvious misuse cases. Fake imagery of terrorist attacks, geopolitical flashpoints, and border crises are among the first things people generate when given the ability to create convincing visual fakes tied to real-world locations.

What to watch

Google says it is "working on implementing stronger guardrails" before potentially re-releasing the feature. Key questions for any future rollout include: Will generated images carry more persistent, hard-to-remove provenance markers? Will Google restrict generation near sensitive locations? Will there be a review or delay before generated content can be screenshotted or shared?

The incident also sets a precedent for how other mapping and geospatial platforms—such as Apple Maps, Mapbox, or open-source projects—approach AI-generated content. If Google Earth, with its reputation for reliability, struggled to manage this, smaller platforms may face even greater challenges.