Google disables AI satellite image feature over disinformation risks
Google briefly launched an AI feature in Google Earth that let users generate photorealistic but fabricated satellite images of real locations. Researchers quickly demonstrated it could create convincing false scenes of wars, disasters, and other scenarios. Google removed the feature within 24 hours and said it would add stronger safeguards.
What happened
Google announced a new feature in Google Earth that allowed users to generate AI-created satellite images overlaid on real locations by entering text prompts.
AI researcher Henk van Ess demonstrated the feature could be used to create realistic-looking false images, including refugees near the Mexican border, a nuclear plant in Iran, bomb craters in Gaza, and other fabricated geopolitical scenarios.
Google disabled the feature within approximately 24 hours of launch, stating it would implement stricter control measures while acknowledging that the public trusts Google Earth to provide a reliable view of the world.
Reported impact
The feature risked creating a readily available tool for generating convincing disinformation, potentially undermining the use of satellite imagery by journalists, researchers, and investigators to verify global events and document crises.
Index analysis
The AI model lacked sufficient content filters and safety barriers to prevent users from generating false geopolitical scenarios, military imagery, and disaster scenes.
Generated images blended realistically with authentic satellite imagery through inherited lighting, shadows, and angles, making them difficult to distinguish from genuine satellite captures.
Governance controls
Google stated it would develop stricter control measures before redeploying the feature.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.