>A new feature allowing users to generate custom, AI-powered images directly within Google Earth has been abruptly withdrawn following a fierce backlash from open-source intelligence (OSINT) experts who decried it as an irresponsible gift to disinformation campaigns.
The tool, announced on Thursday, July 30, was designed to let users “virtually reimagine any place in the real world.” By zooming into a location and typing a description, users could generate a photorealistic satellite-style image, blending AI with Google’s vast library of satellite, aerial, and 3D data. The company suggested it could be used for visualizing historical sites, creating construction plans, or imagining a location’s future appearance.
A “Catastrophic” Blow to Credibility
The feature’s launch was met with immediate and severe criticism. OSINT specialists, whose work relies on verifying open-source content like satellite imagery, quickly demonstrated the tool’s potential for harm. One expert generated a fake satellite image showing a non-existent hospital in the Gaza Strip, complete with a bomb crater nearby. Another created a depiction of a nuclear facility in Iran.
“This is incredibly irresponsible and immediately and catastrophically undermines the credibility of satellite imagery,” one OSINT account posted on X. The sentiment was widely shared, with critics arguing the feature provided an effortless way to manufacture convincing evidence for propaganda and fake news.
Google’s Rapid Response and Unanswered Questions
Facing a growing crisis, Google moved swiftly to defend the tool before ultimately removing it. The company stated it takes disinformation very seriously and that every AI-generated image was embedded with a digital watermark called SynthID. In theory, users could verify an image’s origin by using the Gemini app or Google Lens.
However, this defense quickly crumbled. Users immediately questioned the reliability of SynthID detection, a concern seemingly validated by Google’s own support pages, which acknowledge the technology’s significant limitations. The gap between the tool’s power to deceive and the watermark’s ability to be detected proved too wide a chasm for the public and experts to accept, forcing the tech giant into a humiliating retreat.
