Google has yanked a Google Earth feature roughly 24 hours after switching it on, following a wave of criticism that it made fabricating satellite imagery trivially easy.
The tool, launched Thursday, let anyone edit or generate satellite views using plain text prompts and superimpose the results over real Google Earth maps. The Verge, which illustrated its coverage with a Google image of Pompeii reimagined using Nano Banana 2, described it as a way to "create AI deepfakes of the real world using text prompts."
According to NPR's Geoff Brumfiel, Google said it was rolling the feature back to add "stronger guardrails" after concerns surfaced that it could be used to produce deepfake satellite imagery. TechCrunch reported the retraction came one day after launch amid criticism the tool would spread misinformation. Engadget noted users had less than a day to make misleading Google Earth images before the feature disappeared, and called the additions needless. Ars Technica, summarizing the backlash with the line "What on earth is Google doing?", framed the episode as Google Earth risking its reputation on a tool for making fake satellite pictures.
The speed of the reversal is the notable part. This was not a leaked prototype or a research demo — it shipped to a mapping product that hundreds of millions of people treat as a straightforward record of what the planet actually looks like.
Why it matters: satellite imagery is one of the few forms of visual evidence still widely trusted to settle disputes about wars, disasters, and construction on the ground, and building a one-click fake-generator into the most recognizable consumer map product put that trust at risk faster than Google could defend it.