After people used Nano Banana 2 in Google Earth to make misleading images, Google is rolling it back “while we work on implementing stronger guardrails.”
Google had originally integrated its Nano Banana 2 image model into Google Earth, pitching it as a way to visualize a location’s past or future using natural language prompts. However, users had the ability to ask it to do whatever weird scenario you came up with, too. It was supposed to be a way for students and teachers, for example, to help generate renders for historic cities and places, like searching for the ancient Roman city of Pompeii, and ask Nano Banana to render a hyper-realistic view of what the city looked like in 78 A.D.
Unfortunately, some people took that capability to the dark side.
The company states it has “seen people sharing screenshots of generated imagery that appear to violate our policies.” It noted that “generated images didn’t appear in the main Google Earth experience for others to see and were watermarked as AI-generated.” However, in an era where AI-generated images can be used to push a certain narrative and rile people up online, it was easy to see how bad actors could use the tool irresponsibly.
As Nathaniel Raymond, executive director of the Humanitarian Research Lab at Yale’s School of Public Health, tweeted: “@HRL_YaleSPH deals with attempted disinformation related to overhead imagery and open source info on a near daily basis. Nano Banana opens floodgates to sum of all fears level spoofing of sat data that will hurt vulnerable people.”
Aravind Ravichandran, founder and CEO of TerraWatch Space, also noted that “Google Earth is now the most commonly agreed and used baseline layer by journalists, investigators, and the public,” but the Nano Banana 2 tool means “screenshots of the same trusted interface with fabricated information could be passed on. Think made-up images of people at borders, tanks in towns, burning oil terminals etc.”
Sigh. And this is why we can’t have nice things.
—
Photo Credit: Koshiro K / Shutterstock.com
