Google recently launched a new feature for Google Earth powered by an AI model called Nano Banana 2, which allowed users to modify satellite imagery using simple text prompts. The tool was intended to let users reimagine landscapes, but it was removed just one day after its release following concerns regarding its potential for misuse.

Digital investigator Henk van Ess highlighted the risks of the technology by using the tool to generate manipulated images. Specifically, van Ess created AI-generated scenes depicting refugees near the U.S.-Mexico border and a bomb crater situated next to a hospital in Gaza. These demonstrations illustrated how the tool could be used to create realistic but fabricated visual evidence of real-world events, effectively acting as a deepfake generator for satellite data.

In its initial response to these findings, Google stated that images produced by the Nano Banana 2 model included digital watermarks intended to identify them as AI-generated. However, the company ultimately decided to disable the feature entirely shortly after its debut. While Google has not provided a detailed public statement regarding the permanent future of the tool, the swift removal suggests a prioritization of safety and accuracy over the creative capabilities the feature initially offered.