Google Earth Pulls AI Image Tool After Manipulation Concerns

The controversy began when digital investigator Henk van Ess showcased the platform's vulnerabilities. By inputting specific prompts, he successfully generated fabricated images showing refugees at the Mexican border and bomb craters near a hospital in Gaza. Despite Google's initial defense that the system included digital watermarks and content filters, van Ess proved that the tool could bypass existing safeguards. He further demonstrated that the generated media could successfully deceive AI detection software, such as Hive, rendering the company's internal watermarking efforts ineffective for preventing misinformation.

Google acknowledged the security gaps in a follow-up statement, admitting that users were sharing screenshots that violated core safety policies. While the company maintained that these AI-generated images remained private and were not integrated into the public Google Earth view, the risk of external manipulation proved too high. The feature is now offline while engineers work to implement more robust guardrails. For now, the experiment highlights the difficulty of applying generative AI to platforms that users rely on as sources of objective truth.