Google’s AI Image Generator in Google Earth Lasted Less Than a Day—Here’s Why OSINT Experts Called It a Crisis

Why It Was Pulled

Policy-violating fakes emerged instantly. Users generated convincing fake scenes including a blast crater in Los Angeles, a flooded U.S. Capitol, a fire on Iran's Kharg Island, and—most alarmingly to experts—a fabricated image of a nuclear plant in Iran and Russian tanks rolling through Kyiv .

BBC Verify and OSINT researchers publicly demonstrated the risk. AI researcher Henk van Ess and BBC Verify produced fabricated satellite scenes of sensitive geopolitical locations, showing how easily the tool could generate misleading imagery that appeared authentic .

Google acknowledged the trust problem directly. In its rollback statement, the company said: "We know that people uniquely trust Google Earth for a reliable view of the world. We've seen geospatial professionals using this feature for a range of useful purposes" but that it was "rolling back this feature in Google Earth while we work on implementing stronger guardrails" .

The OSINT Trust Crisis

Satellite imagery has been a gold-standard source for open-source intelligence (OSINT), human rights investigations, and war crime documentation. Experts warned that embedding AI generation directly into Google Earth—the most widely trusted satellite mapping platform—could fundamentally undermine the credibility of all aerial imagery as evidence .

The feature blurred the line between real and fake at the source. Because the AI images were generated within Google Earth itself and tied to real coordinates, they carried a false veneer of authenticity that would be much harder to debunk than a generic AI image shared on social media .

Human rights investigators rely on verified satellite imagery to document atrocities, conflict damage, and environmental crimes. If even Google Earth's own output could be questioned, the entire evidentiary chain for open-source investigations would be damaged, making it easier for bad actors to dismiss real footage as AI-generated .

The OSINT community described the incident as a "trust crisis" for digital evidence. The quick backlash from OSINT professionals was a primary driver of Google's decision to reverse course within hours .

Google's Defense and Its Limits

Google pointed out that all AI-generated images were embedded with SynthID, a digital watermark designed to flag AI-generated content . However, researchers demonstrated that detection tools could be fooled and that screenshots—the most common way fake images circulate—strip away any platform-level watermark or context . The defense was insufficient to prevent the flood of convincing fakes that appeared within hours of launch.

Broader Implications

The incident represents one of the fastest product reversals in Google's history and a case study in the dangers of embedding generative AI into trusted information platforms without adequate safeguards. For the OSINT community, it was a stark warning that the credibility of satellite imagery—a cornerstone of modern digital evidence—could be eroded in a single product launch .