ทันเอไอ

Google removes AI image generation feature from Google Earth just one day after launch

Google has suspended the AI image generation feature on the web version of Google Earth just a day after its launch, following concerns over misinformation after users overlaid fake images onto real coordinates.

Google recently launched an AI image generation feature on the web version of Google Earth, powered by the Nano Banana 2 model, which allowed users to type text prompts to simulate overlays on real coordinates and satellite imagery. However, the feature quickly ran into trouble and was suspended just one day after release.

The reason was that researchers and users found the feature being used to create realistic fake images (deepfakes) overlaid on real locations, such as simulated war craters, nuclear power plants, or nonexistent disaster zones. This sparked concerns that these images could be spread as misinformation and undermine Google Earth's credibility as a reliable source of satellite imagery used by people and organizations worldwide.

Google clarified that some users were found to have created images violating usage policies, prompting a temporary shutdown to develop stricter safety systems and guardrails. The company confirmed that all generated images were embedded with SynthID watermarks (identifying them as AI-generated) and that these images were not integrated into Google Earth's actual map database.

This incident highlights a design challenge faced by many generative AI tools: when linked with trusted real-world data, the ability to 'simulate' can easily become a channel for creating deceptive images indistinguishable from reality.