[REVISION HISTORY]

Google Earth AI image tool withdrawn after backlash

Updated 31 times since CLSTR started tracking revisions of this situation.

What changed

2026-08-06 23:04 UTC → 2026-08-09 06:54 UTC · added ~~removed~~

Google launched the Nano Banana 2 “Create Image” AI feature in Google Earth on 30 July 2026, letting users generate photorealistic overlays on satellite, aerial and 3‑D map layers. Within hours researchers and open‑source intelligence experts demonstrated that the tool could fabricate nuclear facilities, graphic scenes of suffering in Gaza, a bomb crater in Los Angeles, protestors outside Google’s Mountain View campus, a fake nuclear plant in Iran and a simulated refugee camp on the US‑Mexico border. The images bore a SynthID watermark and were excluded from the main Google Earth experience, yet screenshots spread on social media. Citing policy violations and the risk of weaponising the technology for disinformation, Google disabled the feature on 31 July, less than 24 hours after launch, and announced a full rollback on 2 August while stronger guardrails are built. A company statement noted that geospatial professionals found the feature useful, but misuse forced a formal suspension on 3 August and warned that the tool could be abused to fabricate realistic images of nuclear sites, refugee camps and conflict zones. Experts highlighted that even watermarked AI‑generated images could be stripped and shared, undermining the credibility of geospatial data, and said the risk outweighed the benefits. The episode underscores ongoing challenges in policing user‑generated AI content in critical visual domains.

Versions

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.