A recent experiment by Digital Digging highlighted the dangers of integrating an AI image generator into Google Earth. Henk van Ess demonstrated how believable flyover images could be created, potentially spreading misinformation.

Google’s initial response was to emphasize its SynthID digital watermark and suggest users employ additional verification methods such as comparing imagery from other satellite platforms or checking for suspicious tags in Gemini.

The feature launched globally on Google Earth's web version, pitched as a tool for visualizing historical sites and real estate projects. However, the ease with which it can generate convincing images raises concerns about its potential misuse.

While Google’s watermark makes detection difficult but not impossible, it failed to stop Digital Digging from fooling Hive’s AI detector with an altered video from Google Earth. This incident underscores the need for robust measures against disinformation in tech tools.

Update: Google has since rolled back the feature to prevent further misuse and misinformation.