Google Removes Earth AI Image Tool Over Deepfake Risks
Google has removed an AI image generation feature from Google Earth just one day after its launch. The decision followed strong criticism from journalists, security researchers, and open-source investigators. They warned that the tool could increase the risk of geospatial misinformation by creating realistic but false satellite images.
The feature used Google’s Nano Banana 2 AI model and allowed users to add AI-generated images over real satellite maps through text prompts. Initially, the tool was designed for creative purposes, such as visualizing historical events, planning real estate projects, and imagining future locations.
However, researchers quickly showed how easily the feature could create misleading images of sensitive situations. For example, Henk van Ess, an open-source researcher and founder of the investigative publication Digital Digging, created fake images involving refugees at the Mexican border, a nuclear facility in Iran, and a hospital with a bomb crater in Gaza. “Nothing was refused, nothing was softened, and nothing suggested I try a different prompt,” van Ess told NPR.
Additionally, NPR generated artificial images showing Iran’s Kharg Island on fire and a flooded U.S. Capitol complex. Meanwhile, Evan Hill, a visual forensics investigator at the Washington Post, wrote on X that he was “very curious how or if this idea was red teamed internally because the opportunities for abuse and disinfo are literally boundless.”
AI Tool Removed After Policy Concerns
On Friday, Google acknowledged the concerns and announced that it was rolling back the feature. “We know that people uniquely trust Google Earth for a reliable view of the world,” the company said. “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.”
Furthermore, Google explained that generated images contained its SynthID digital watermark and were not visible in the main Google Earth experience for other users. However, critics argued that these safeguards were not enough to prevent misuse.
For instance, van Ess tested his fake satellite imagery through the AI detection tool Hive. Surprisingly, the system rated the images as 99.2% likely to be real, raising further concerns about the effectiveness of current detection methods.
The Importance of Trustworthy Satellite Images
The controversy highlighted a larger issue because satellite imagery has become an important verification resource for journalists, researchers, and human rights investigators. Moreover, these images are often used to confirm events in conflict zones and locations that are difficult to access.
“Satellite imagery has been kind of a safe bet when it comes to verifying an event because it’s hard to fake,” said Jake Godin, a senior researcher at Bellingcat.
Additionally, van Ess summarized the concern in a statement from his newsletter: “Google spent twenty years building the reference the world checks against. Today it added a button that makes things up.”
Therefore, the incident shows the growing challenge of balancing creative AI innovation with the need to protect public trust in digital information. As AI-generated content becomes more realistic, stronger verification systems and safety measures will become increasingly important.
