Google briefly rolled out an AI image generation feature inside Google Earth, then removed it the following day after users online quickly revealed the problems with letting prompts reshape what appears on the platform.
The reported update was part of Google's wider push to add generative AI tools to its products. But Google Earth is not just another creative app: it is closely tied to real places, familiar landscapes and a basic expectation that what people see relates to the world as it is.
That seems to be where the experiment ran into trouble. Once internet users began testing the tool, the drawbacks of allowing broad prompt-based changes to planetary imagery became hard to ignore, and Google moved fast to take the feature back down.
The short lifespan of the launch highlights a broader issue facing AI rollouts. Generative features can look appealing in demos, but when they are added to products built around geography, accuracy and public trust, the margin for error becomes much smaller.
