Google pulls Earth AI image feature after 1 day amid misinformation fears
Nano Banana 2 hit Google Earth, then vanished within 24 hours.
Google's brief experiment with injecting AI image generation into Google Earth lasted barely a day. On Thursday, the company debuted Nano Banana 2, its latest prompt-based image model, inside the satellite mapping app, inviting users to "get creative with geography" by superimposing virtually any generated image onto real-world terrain. The feature effectively turned one of journalism's most trusted visual evidence sources into a canvas for fabrication, and critics immediately flagged the risk.
BBC journalist Shayan Sardarizadeh posted a sarcastic warning, noting the tool "could possibly be abused to spread misinformation online." By Friday, Google reversed course, stating it had seen useful applications from geospatial professionals but also screenshots "that appear to violate our policies." The company is rolling the feature back while it adds stronger guardrails—a tacit admission that even its own product could become a disinformation engine.
- Nano Banana 2 was integrated into Google Earth and pulled within 24 hours.
- Critics included BBC journalists who warned of geospatial disinformation from prompt-based map editing.
- Google cited policy-violating generated screenshots and is implementing stronger guardrails before any relaunch.
Why It Matters
Trusted visual evidence tools are becoming vulnerable to AI fabrication, forcing platforms to choose between creative features and information integrity.