Google Earth Pulls AI Image Feature in Under 24 Hours After Deepfake Disasters
jonerp · x · 2026-08-02
Google Earth recently launched an AI image generation feature powered by Gemini's "Nano Banana 2" model, allowing users to alter map locations with text prompts. However, the feature was pulled in less than 24 hours.
The rollback occurred after users quickly exploited it to create deepfake disaster imagery. AI researcher Henk van Ess demonstrated generating fake refugee camps along the US-Mexico border and a nuclear plant in Iran. Others created images of fake flooding and 9/11-style attacks. These highly realistic fake disaster images alarmed emergency managers, forcing Google to act swiftly.
More from Fun
- At least one industry hasn't forgotten us, hinting at AI industry attention — armano · 2026-08-02
- Codex Agent Scores Perfect 38-0 in Soccer Simulation Game — talkaboutdesign · 2026-08-02
- AI figures tease 'AGI is upon us' in St Tropez photo — teodorio · 2026-08-02
- AI Coding Debate: Devs Refuse to Read Generated Code for Ego? — dotey · 2026-08-02
- Developer Confesses to Constantly Rebuilding AI Agent Setups, Only to Return to Cursor — brandon_galang · 2026-08-02
- Dev Jokes About Switching Side Projects Weekly from Codex's POV — haydendevs · 2026-08-02