Google Scraps Google Earth’s AI-Generated Imagery Feature After Misinformation Warnings

Google’s latest experiment with AI inside Google Earth barely had time to find an audience before it was pulled. The feature promised a fast way to generate satellite-style imagery for everything from historical visualization to real estate mockups, but researchers and open-source investigators quickly flagged a darker upside: convincing fakes that could travel faster than any correction. With trust in geospatial evidence already under strain, critics argued the tool risked turning a widely used reference map into a new vector for misinformation.

Google halts its latest AI experiment

If you blinked, you missed it. This week, Google quietly added a new AI feature to Google Earth, then pulled it back down in less than 24 hours after researchers raised alarms about misinformation. The concern was simple: when a mapping product becomes a trusted reference, even a small “make it up” button can bend the internet’s sense of reality.

Google confirmed the reversal in a statement, saying it had seen people share screenshots of AI-generated images that “appear to violate our policies,” and that the company is pausing the feature while it builds stronger safeguards. The tool is no longer visible inside Google Earth.

What the tool actually did

The feature, labeled “create an image”, let users zoom into a location and generate fresh satellite-style imagery in seconds, using Google Earth’s underlying satellite, aerial, and 3D mapping data. Google pitched it as useful for benign work, for example visualizing how a landscape looked in the past or sketching scenarios for urban development and real estate planning.

On paper, that sounds like a neat creative layer on top of a familiar product. In practice, the output could look close enough to authentic imagery that it risked being passed around as evidence, especially on fast-moving social platforms where context gets stripped away.

Why researchers called it a misinformation risk

The sharpest criticism wasn’t that AI images can be fake, we already know that. It was that fabricated satellite imagery hits a particularly sensitive nerve: geospatial screenshots are a staple for open-source investigators, journalists, and everyday fact-checkers trying to verify events remotely.

Digital investigations expert Henk van Ess wrote that Google spent 20 years building a reference the world relies on, then added a button that invents things. Brady Africk, a satellite imagery analyst at the American Enterprise Institute, warned that convincing fake satellite images could spread quickly and “erode public trust” in real imagery, making verification harder.

Tests reported by journalists showed the tool could be pushed toward harmful scenarios, including depictions tied to sensitive military or geopolitical locations. That echoes recent history: in 2023, an AI-generated image of an “explosion” near the Pentagon briefly rattled markets.

Google’s safeguards, and what comes next

Before the pause, Google said AI-generated images included invisible digital markers that users could detect with Gemini, and that it blocked generation on “dangerous” topics under its AI use policies. Critics argued that enforcement and detection are only as strong as the real-world testing, and that screenshots travel farther than provenance tools.

Google says the feature may return with tougher guardrails. The bigger question is whether geospatial truth, once softened by generative AI, can be reliably re-hardened for the people who need it most.