Google Pulls AI Deepfake Map Tool After 24 Hours
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Google launched a new AI feature in Google Earth on Thursday letting users type "type whatever you want to see" to generate synthetic aerial images, available to billions of users worldwide.
- Google disabled the feature on Friday, roughly 24 hours after launch, citing screenshots of generated imagery that appeared to violate its policies.
- Open-source intelligence researchers demonstrated within that window that the tool could produce realistic-looking nuclear power plants, refugee camps along borders, burned forests, flooded cities, and fatal car crashes.
- Google intended the feature for imaginative but benign uses like visualizing empty lots, historical neighborhood views, and speculative future landscapes, according to the article.
- Google Earth has served for 20 years as a verification and debunking tool for journalists, governments, and investigators — a role that the AI feature directly threatened, the source notes.
- Google stated that geospatial professionals had also used the feature for "useful purposes," framing the shutdown as a policy-enforcement decision rather than a full retraction of the technology.
Why it matters: The 24-hour window exposed how a single open-ended prompt can turn a trusted verification tool into a disinformation factory, giving anyone the ability to fabricate location-specific imagery that looks like satellite evidence. For OSINT analysts and journalists who have relied on Google Earth for two decades to debunk falsehoods, the episode raises the stakes on distinguishing real captures from AI-generated ones.
