California (GNP): Google has disabled a newly launched artificial intelligence feature on Google Earth after it was found capable of generating realistic deepfake satellite imagery, just a day after the tool was introduced.
The feature, which allowed users to type simple instructions to generate AI-created aerial visuals, was intended to let people imagine alternate scenes, such as visualizing an empty lot with new development or picturing how a neighborhood might have looked a century ago or could look decades into the future.
For roughly 24 hours after its launch on Thursday, the tool instead enabled users to generate convincing fabricated imagery of scenes that never occurred, including fictitious nuclear power plants, refugee camps placed along real borders, burned forests, flooded cities and depictions of a fatal car crash.
Open-source intelligence researchers and journalists were among those who demonstrated how easily the feature could be used to produce such content, raising concern given Google Earth’s two-decade role as a widely used tool for verifying real-world events and countering false claims.
Also Read: OpenAI Finds More AI Agents Escaped Containment
Google said in a statement that while some geospatial professionals had used the feature for legitimate purposes, it had also observed people sharing screenshots of generated imagery that appeared to violate company policy. Citing those concerns, the company said it was disabling the tool just a day after it had publicly introduced it.
