Google Launched an AI Image Generator in Google Earth — and Rolled It Back in Less Than 48 Hours

Immediate backlash and demonstrations of harm

Within hours of launch, security researchers and journalists began demonstrating why the tool was dangerous.

Henk van Ess of Digital Digging generated fabricated satellite images showing refugees at the U.S.-Mexico border and bomb craters near a hospital in Gaza — scenes that, without context, looked like authentic satellite evidence . BBC Verify created images of a collapsed Eiffel Tower, a sinkhole swallowing the Great Pyramid of Giza, and Russian tanks in Kyiv . Other users generated fakes of explosions, nuclear sites, and militant bases anchored to real geographic coordinates .

As one researcher put it, the tool made it trivially easy to create "deepfakes of the real world" .

Key criticisms

Erosion of trust in satellite imagery: Journalists, open-source intelligence (OSINT) analysts, and humanitarian researchers said the feature threatened the foundational credibility of satellite imagery, which is routinely used to verify war crimes, disasters, and human rights abuses . Multiple commentators noted that Google Earth has long been treated as authoritative ground truth, and embedding generative AI directly into that interface created a dangerous "what's real anymore?" dynamic .

Inadequate safeguards: Google said AI-generated images included invisible watermarks, but BBC Verify found those safeguards were easily bypassed and that third-party detection tools could not reliably flag the fakes . Critics argued that by the time a watermark was checked, the image could already have spread virally across social media stripped of metadata .

Geospatial misinformation risk: The tool allowed users to place fabricated events — bomb craters, refugee camps, military bases — onto real coordinates, making the fakes look like credible satellite evidence . Multiple outlets warned this could fuel real-world conflicts and disinformation campaigns .

Google's response and rollback

On July 31, 2026, Google announced on X (formerly Twitter) that it was pausing the feature. The company stated:

"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. So we're rolling back this feature in Google Earth while we work on implementing stronger guardrails."

The company did not announce a timeline for the feature's return .