Relevant NewsDaily Edition
Google has disabled an AI feature it launched in Google Earth that allowed users to generate satellite imagery edits using text prompts, after researchers demonstrated the tool could create convincing deepfakes depicting false crises like refugee camps and bomb sites. The tool, powered by a model called Nano Banana 2, included watermarks and content filters but proved insufficient to prevent misuse within 24 hours of launch.
Why it matters: The rapid shutdown highlights the tension between AI innovation and real-world harm prevention—a critical challenge as generative tools become more accessible and capable of producing geopolitical disinformation at scale.
