Google Enhances Earth Tool to Combat Deepfake Satellite Risks
WHO:
NPR - Google, Bryan Horowitz (product manager), Henk van Ess (open-source researcher), Jake Godin (senior researcher with Bellingcat), Evan Hill (visual forensics investigator at the Washington Post).
WHAT:
NPR - Google announced it is rolling back the AI image generation tool in Google Earth to implement "stronger guardrails" after concerns about its potential misuse for creating deepfake satellite imagery.
NPR - Google stated, "However we've also seen people sharing screenshots of generated imagery that appear to violate our policies," prompting the rollback of the feature.
NPR - Google rolled back an image generation tool in Google Earth designed to create AI-generated satellite images.
NPR - The tool was removed after concerns that it could be used to create deepfake satellite imagery.
NPR - Google stated, "We've seen geospatial professionals using this feature for a range of useful purposes," but also noted that "people [were] sharing screenshots of generated imagery that appear to violate our policies."
NPR - The initial announcement of the feature allowed users to "zoom in to a place in Google Earth on web, tap 'create image,' and type whatever you want to see."
NPR - NPR was able to generate images of events that had not occurred, such as Iran's Kharg Island on fire and a flooded U.S. Capitol complex.
WHERE:
NPR - Google Earth platform, locations mentioned include Iran's Kharg Island and the U.S. Capitol complex.
WHY:
NPR - Experts expressed horror at the potential for the tool to create deepfake satellite imagery at the click of a button, with Henk van Ess stating, "I tried refugees at the Mexican border, a nuclear plant in Iran, a crash in Amsterdam, a hospital with a bomb crater in Gaza. Nothing was refused."
NPR - Concerns arose regarding the potential for creating deepfake satellite imagery that could mislead the public and undermine trust in real satellite images.
NPR - Jake Godin expressed that the tool could "vastly accelerate the generation of AI fakes," leading to increased misinformation.
HOW:
NPR - Google indicated that it will prevent image creation on harmful topics and is continually updating its protections against misuse of the tool.
NPR - The tool allowed users to create AI-generated satellite images with a single click, streamlining the process of generating fake images.
NPR - Google stated that all images generated would be marked with a SynthID watermark to indicate they are AI-generated.
(UPDATED July 31, 2026 20:22 UTC)
Google has announced it is rolling back its AI image generation tool in Google Earth to implement stronger guardrails after concerns about potential misuse for creating deepfake satellite imagery. The decision comes after experts raised alarms about the tool's capabilities, with one expert highlighting that he was able to generate images of sensitive locations without any restrictions. Moving forward, Google plans to prevent image creation on harmful topics and will continue to update its protections against misuse of the tool.
--
(UPDATED July 31, 2026 19:25 UTC)
Google has decided to roll back an image generation tool in Google Earth that allowed users to create AI-generated satellite images, following concerns about its potential to produce misleading deepfake imagery. The tool, which enabled users to generate images of events that had not occurred, raised alarms among experts like Jake Godin, who warned it could "vastly accelerate the generation of AI fakes." While Google acknowledged that the feature was being used for legitimate purposes by geospatial professionals, they also noted that some generated images appeared to violate their policies, prompting the removal of the tool.