Google has decided to disable its new feature in Google Earth that allowed users to create AI-generated visuals over satellite images, due to strong criticism from researchers and open-source intelligence professionals concerned about disinformation risks.
A Google representative told AFP, “While geospatial experts have found some useful applications for this feature, we’ve also seen screenshots of generated images that breach our policies, so we’re rolling back the feature for now as we develop better safeguards.”
They added, “We understand that users rely on Google Earth for an accurate view of the world.”
Launched on Thursday, the “create image” tool integrated Google’s Nano Banana 2 image-generation tech into Google Earth, enabling users to zoom into locations and craft images quickly based on satellite, aerial, and 3D mapping data.
Google highlighted its potential for visualizing historical events, planning real estate projects, and more. However, critics warned that this capability could be misused to produce highly realistic geospatial fakes, undermining trust in satellite imagery—a crucial tool for fact-checking and verification.
Henk Van Ess, an expert in digital investigations, remarked on Substack, “Google has spent two decades building a trusted reference for the world, and now they’ve added a button that simply fabricates images.”
Brady Africk, a satellite imagery analyst at the American Enterprise Institute, told AFP that the new feature made it “easier to generate convincing fake satellite images that could spread rapidly online and mislead the public.” Such fakes could erode confidence in satellite data and complicate the work of journalists and researchers.
In response to mounting concerns, Google stated on X that images generated with the feature included SynthID digital watermarks—hidden markers embedded into AI-created media to help identify authentic content.
AFP’s testing revealed that with this tool, one could create images depicting an explosion in Paris, a nuclear facility in Iran, a bomb crater in Russia, or an ISIS training camp in Syria, all of which could have serious global implications. The tool also produced fake scenes of flooding in Bangladesh, a migrant caravan near the U.S.-Mexico border, and a suspicious warehouse in Georgia, a state where former President Donald Trump alleged widespread election fraud.
Following the backlash, Google removed the “create image” button from Google Earth.
Past incidents involving AI-generated images have already caused real-world issues—such as briefly rattling financial markets in 2023 with a fabricated Pentagon explosion, or gaining millions of views for a false depiction of a destroyed U.S. military base early in the Iran conflict.
Organizations like the London-based Centre for Information Resilience have called for the reversal, emphasizing that trust in satellite imagery, built over years, could be lost overnight. GeoConfirmed, an open-source project monitoring conflict visuals worldwide, noted that malicious actors had already begun exploiting the tool to map military positions and alter building labels, recognizing that its primary use appears to be the creation and spread of misinformation.