Google Earth Fakes? Trust Just Shattered

By briefly letting anyone fake satellite scenes with a text prompt, Google Earth has shaken trust in one of the world’s most important “eyes on the truth.”

Story Snapshot

Google Earth added an AI “create image” tool that let users drop fake events onto real satellite views.
The tool used Google’s Nano Banana 2 image model and went live worldwide on July 30 before being pulled back.
Researchers, journalists, and fact-checkers warned it could turn a core verification tool into a source of confusion.
Google paused the feature after backlash, but the episode raises deeper questions about trust in big tech and digital maps.

Google Earth Turns On, Then Quickly Pauses, AI Satellite Tool

Google Earth, the popular mapping site many people use to see real satellite views of the world, quietly added a new “create image” button that did something very different from what users expected maps to do. With this tool, anyone could zoom to a real location, type a short description, and ask Google’s Nano Banana 2 image generator to build a synthetic picture right on top of actual satellite, aerial, or 3D map data. The feature went live on the web version of Google Earth around July 30 and rolled out globally within hours.

Google promoted the tool as a way to “transform any place” and “bring your ideas to life,” saying people could visualize history, mock up future buildings, or design real estate plans by layering imagined scenes onto real coordinates. The company’s blog and press language stressed creativity and said Nano Banana 2 would use real-world knowledge to ground concepts in existing terrain and city layouts. For many everyday users, the idea sounded fun, like turning the planet into a digital sketchbook instead of a static atlas.

Experts Show How Easy It Is to Fake War Zones and Disasters

Almost immediately, open-source investigators and journalists started testing how far the tool could go. One reporter described generating violent scenes, including bomb damage and armed clashes, over well-known locations by simply clicking “create image” and entering a few words in the prompt box. Another investigator showed images of “refugees near the Mexican border” and a bomb crater next to a hospital, all built on top of real satellite imagery so that the angle, shadows, and surroundings looked convincing to casual viewers. These were not harmless cartoons; they looked like real evidence of serious events.

Researchers who track disinformation warned that the tool could be a gift to people who want to stir fear or anger online. Google Earth is widely used by journalists, human-rights groups, and citizens to verify where something happened, from a protest march to a missile strike. When fake images are created inside the same platform, with real coordinates and realistic details, it becomes much harder for people to tell what is true. Critics said this does not just make new fake pictures; it attacks the basic trust in satellite imagery as a neutral record we can all rely on.

Google’s Guardrails: Watermarks and Blocks, But Extra Steps for Users

Facing questions, Google stressed that it “takes misinformation seriously” and that every image made with Nano Banana in Google Earth carries a SynthID digital watermark. That watermark is invisible to the eye and is meant to mark the content as AI-generated so that other tools, like the Gemini chatbot app or Google Lens in Search, can spot it later. Google also said the generated pictures did not appear in the default Earth view for everyone, but were saved inside a user’s project space. In theory, these steps should help keep synthetic scenes separate from standard satellite layers.

Critics pointed out a practical problem: regular people on social media do not run every picture through a verification tool. If someone downloads a dramatic AI satellite image from their Earth project and posts it with a scary caption, many viewers will react based on the image alone. Only highly trained users, such as researchers or journalists, are likely to compare the scene with other satellite services or check orbital data to see if it is real. That means bad actors can move faster than the average citizen’s ability to confirm the truth.

Roll-Back After Backlash and What It Says About Big Tech Trust

Within about a day of launch, as fake satellite scenes began spreading and experts raised alarms, Google announced on X that it was pausing the new Earth feature “while we work on implementing stronger guardrails.” Reports from several outlets confirmed that the “create image” button had disappeared and the Nano Banana 2 integration was no longer accessible to users. Google told one news agency that it had also seen people sharing screenshots from the tool and suggested that this behavior helped drive its decision to pull back.

What breaks when AI can fake any place on Earth?

The issue was not that Google Earth could generate fake images. It was that it could make fake images inherit Google Earth’s authority in 24 hours.

On July 30, Google launched Nano Banana 2 inside Earth: select any coordinates,…

— Nox Vector (@NoxVectorAI) July 31, 2026

This short-lived experiment taps into a larger frustration shared by many Americans, both conservative and liberal: powerful tech companies keep rolling out tools that can reshape our information environment, but ordinary citizens carry the risk when things go wrong. Maps and satellite views are supposed to be boring and trustworthy—part of the common ground people use to challenge false claims from governments, corporations, and even the media. When synthetic images sit one click away inside those same tools, the line between fact and fantasy blurs just enough to help those who want to mislead.

For Americans worried about wars, immigration, energy sites, and protests, the stakes are real. A single fake “satellite photo” of a border clash or a pipeline explosion could race across social networks and cable news before anyone checks if it came from an AI prompt. When that fake carries Google Earth’s look and feel, it borrows the trust the platform earned over years of showing the world as it is. Even though Google has now paused the feature and talks about better guardrails, the episode is a reminder that the gatekeepers of our digital reality are willing to test risky ideas first and think about consequences later.