Google Pulls Earth AI Less Than a Day After Launch Over Disinformation Risk

On July 31, 2026, Google disabled the Earth AI feature less than a day after launch. The tool let any user generate synthetic images and overlay them on real Google Earth satellite imagery. It immediately drew a wave of criticism over the obvious risk of spreading disinformation.

How Earth AI worked

Earth AI let users create AI-generated images and embed them directly into the real maps of Google Earth, one of the world's most widely used mapping tools, with detailed satellite imagery. The technical detail that set off alarms: the resulting image was never labeled as synthetic. An artificial object blended seamlessly into a genuine satellite image, with no warnings or watermarks.

According to TechCrunch, access to the tool was open to a broad group of users, with no restrictions on the type of content they could overlay. In practice, anyone could "place" a nonexistent building, vehicle, or any other structure on the map of any location in the world, and get an image visually indistinguishable from a real satellite photo.

Key facts about the launch and shutdown:

• Earth AI launched on July 30, 2026
• It was disabled on July 31, 2026, less than 24 hours after release
• Access was broad, with no restrictions on the type of overlaid content
• Synthetic content was not labeled at all
• Reason for the shutdown: a wave of criticism over disinformation risk

Why did the criticism erupt so fast?

The problem was obvious within the first hours of release. Earth AI provided ready-made infrastructure for producing convincing "satellite evidence" of events that never happened: destruction, new facilities, troop or equipment movements. With no labeling and no restrictions, such images could easily be passed off as authentic open-source imagery.

The absence of any mechanism to distinguish real from synthetic content was the main driver of the mounting criticism. The public traditionally treats satellite imagery as an especially credible type of evidence. Journalists, independent investigators, and participants in public debates all rely on it. A tool without safeguards opened the door to manipulating geodata at a scale previously out of reach for an ordinary user without specialized technical skills.

"The tool, which let anyone generate fake AI images and overlay them on real Google Earth maps, quickly drew a negative reaction,"

TechCrunch writes.

As TechCrunch reports, this rapidly growing flood of criticism was the direct reason for the shutdown. Google deactivated the feature in under 24 hours, without waiting for the backlash to grow further. The speed of the reaction is telling: either an internal risk assessment was not completed before release, or the volume of criticism exceeded expectations.

What this means

The Earth AI episode raises a pointed question for the whole tech industry: how should companies assess the risks of AI tools that work with real geospatial data before releasing them to the public? Geolocation content is an especially sensitive area. Manipulated satellite imagery is already used in disinformation campaigns, and a tool that reduces creating it to a few clicks for a mass audience multiplies the potential scale of the threat.

Google responded quickly, pulling the feature less than 24 hours after release. Still, the fact that a tool with such obvious risks reached public launch without safeguards (no labeling, no restrictions, no warnings for users) points to gaps in existing pre-release review procedures. Incidents like this are likely to increase pressure on major tech companies over the responsible release of AI products that have access to real-world data.