Google is quickly rolling back its new AI image generation feature within Google Earth, less than a day after its July 30th announcement. The company cited concerns over the tool being used to create misleading satellite imagery, which violated its policies. While acknowledging its utility for geospatial professionals, Google stated that screenshots of generated images had surfaced, prompting the rollback.

This decision follows a report from 404 Media demonstrating how the feature, powered by Nano Banana, could be easily manipulated. Journalist Joseph Cox used the tool to generate fabricated images, including a blast crater in Los Angeles and protestors near Google's Mountain View campus. Google initially defended the feature, pointing to its Synth ID digital watermark and Gemini's ability to verify AI-generated content.

However, critics argued the watermark was insufficient against the spread of misinformation. Google reiterated its defense of the watermark but also pledged to implement stronger guardrails before re-enabling image generation in Google Earth.