Google removed an artificial intelligence image generator from Google Earth just one day after introducing the feature. The tool, powered by the Nano Banana model, allowed users to alter locations and create fictional events directly on the map. The company said the feature was withdrawn because people were generating and sharing content that violated the platform’s rules.

The experimental service could transform almost any place shown in Google Earth through a simple text instruction. Users were able to add buildings, natural disasters, crowds or other events to real geographical locations, producing images that could appear convincing without additional context.

Several fabricated scenes quickly circulated online after the launch. Among them were images depicting a large explosion crater in Los Angeles and the United States Capitol surrounded by floodwater. Although the pictures were generated artificially, their realistic appearance raised concerns that they could be mistaken for genuine satellite or news imagery.

Open-source intelligence researcher Henk van Ess also tested the limits of the system. According to him, the generator accepted potentially sensitive prompts, including requests to depict refugees near the United States-Mexico border and a nuclear facility in Iran. The tool reportedly produced the requested scenes without blocking the instructions.

The incident demonstrated how quickly generative AI features connected to real maps can be used to create misleading visual content. When fictional events are placed in recognizable locations, viewers may find it difficult to distinguish experimentation from evidence of an actual incident.

Google responded by removing the generator rather than leaving it available while introducing additional restrictions. The company did not specify when or whether the feature might return, but the rapid shutdown suggests that its existing safety mechanisms were not sufficient to prevent misuse.

The short-lived experiment also highlighted a broader challenge for technology companies developing AI-powered creative tools. Such services can offer new ways to explore maps and visualize ideas, yet they also require effective safeguards against the creation of deceptive images involving disasters, conflicts, infrastructure and politically sensitive locations.