Google rolls back a newly launched AI feature in Google Earth after researchers warn it can be used to create misleading, AI-generated satellite scenes. The tool lets users generate images from text prompts and place them on top of real satellite map coordinates, effectively combining invented content with genuine geospatial locations.
Across outlets, experts say the main risk is misinformation: fabricated visuals could be presented as evidence in investigations of events and conflict damage, including incidents on humanitarian and military topics. The BBC reports that while Google says generated images include invisible watermarks detectable with its own tools, testers can still fool some detection methods and produce convincing results. Bloomberg and The New York Times describe the rapid pullback following public backlash, with Google citing concerns that some users were sharing altered imagery that appeared to violate its policies.
Coverage also highlights examples of potential misuse raised by researchers and investigators, including fake disaster imagery and other fabricated scenes tied to real coordinates. Google says it takes misinformation seriously and is updating safeguards while pausing the feature, according to multiple reports.
