Google Earth’s AI deepfake tool only lasted one day

Google launched a prompt-based AI image-editing tool in Google Earth, then rolled it back within a day after users shared satellite images that appeared to break company policy.

The feature was designed for useful geospatial work, but researchers showed it could be abused to fabricate scenes and evade detection systems.

Although the outputs were AI-generated and watermarked, the incident underscored the risks of applying generative AI to trusted satellite imagery.

The episode highlights how altered geospatial scenes can be mistaken for real evidence, making verification and moderation critical.