Google Earth’s AI deepfake tool lasted one day

Google has pulled a Google Earth feature just one day after launching it, following tests that showed users could generate convincing alterations to satellite imagery with text prompts.

The feature used Nano Banana 2 to modify scenes in Google Earth. Digital Digging researcher Henk van Ess deliberately generated images depicting refugees near the Mexican border and a bomb crater beside a hospital in Gaza. He said his testing produced no meaningful resistance from the system:

“Nothing was refused, nothing was softened, and nothing suggested I try a different prompt.”

Recommended reading

AI companies are buying books by the thousands

Google initially told van Ess on Thursday that generated images carried a digital watermark and that the company blocked “image creation on harmful topics.” But the company reversed course the following day after screenshots of generated imagery appeared to violate its policies.

“We know that people uniquely trust Google Earth for a reliable view of the world. We’ve seen geospatial professionals using this feature for a range of useful purposes, however we’ve also seen people sharing screenshots of generated imagery that appear to violate our policies. So we’re rolling back this feature in Google Earth while we work on implementing stronger guardrails.”

Van Ess also reported that a video generated in Google Earth fooled Hive’s AI detection tool, demonstrating how the media could potentially circulate despite its watermark. Google said the generated images did not appear in the main Google Earth experience for other users and were labeled as AI-generated. The company did not say when, or whether, the feature would return after stronger safeguards are implemented.