Overview
- Google added Nano Banana 2 image generation to Google Earth’s web interface on July 30 and removed the feature about 24 hours later after public tests showed easy creation of false scenes.
- Researchers and open‑source investigators used simple text prompts to generate fabricated images such as refugees at the US–Mexico border, a nuclear plant in Iran, bomb craters at a hospital in Gaza and other geopolitical or disaster scenes.
- Google said every generated image carried an invisible SynthID watermark and that it blocks harmful topics, but testers found the watermark can be lost or unreadable when images are shared as screenshots, re‑encodes or phone photos.
- Experts warned the tool risked eroding trust in satellite imagery that journalists and investigators use to verify events, saying one‑click fakes could spread faster than corrections and could be used to dismiss real evidence.
- Google says it will only restore the feature after adding stronger safeguards, but it gave no timeline, leaving questions about what technical or policy changes will be needed to prevent practical misuse.