Edens.nl: het laatste nieuws het eerst! 
Last week, Google rolled out its AI image generation model, Nano Banana 2, to Google Earth, its world explorer program. The company advertised the integration as a way for people to create infographics about different locations around the globe, or to generate historical landscapes of modern-day imagery. Users were also invited to generate their "dream homes" against real-world backdrops—so if you had plans to build an extension on your home, or visions of a vacation house on the lake, you could visualize it with Nano Banana 2 in Google Earth.
The news followed a trend of Google integrating its AI platforms into just about every one of its products and services (for example, last week it added new Gemini capabilities to Google Docs for Workspace users). From the company's perspective, it likely made total sense to add Nano Banana 2 to Google Earth. As it turned out, however, this is one product that Gemini wasn't meant for.
Google Earth has already dropped Nano Banana 2
Just one day after Google launched this big Nano Banana 2 integration, it rolled it back. The company issued an update to its original announcement, which read: "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. It's important to note that generated images didn’t appear in the main Google Earth experience for others to see and were watermarked as AI-generated."
This statement doesn't properly describe the images that users were generating with this tool. In my view, it went beyond "violating" Google's policies: The internet, predictably, used Nano Banana 2 to generate inappropriate and insensitive images based on real locations. Perhaps it won't come as a surprise to learn that one of the images that took off most was of a plane crashing into the new World Trade Center in Manhattan. Other images depicted the Eiffel Tower destroyed, or the Egyptian pyramids collapsing into a sinkhole. While many would likely understand these were produced in jest, plenty more could see these "official" Google Earth images online and assume the worst.
But beyond the jokes, the potential for misinformation and disinformation was alarming. For example, one user demonstrated how they used the tool to place refugees shoulder to shoulder near the Mexican border. One could imagine such an image going viral, with viewers tricked into thinking the Google Earth still was an actual satellite image. Google likely has some guardrails in place preventing many image generations from taking place, but clearly those limitations weren't strong enough. And since anyone, whether they paid for Google's AI features or not, could use Nano Banana 2 with Google Earth, all one had to do was open Google Earth and fiddle with a malicious prompt until Nano Banana 2 complied.
Generative AI has no place in a tool like Google Earth
It doesn't take an AI skeptic to explain why this was a bad idea in the first place—but then again, few AI companies seem to be employing any skepticism before jamming generative AI features into their products. Tools like Google Earth are meant to be educational. You use them to learn about the world, its geography, and what different cities and landscapes look like. I personally love spending time spinning the digital globe around to learn about a new location, and often drop a pin on a random spot to see what it looks like on Street View. The reason I do this is that I understand Google Earth to be an accurate, reliable look at the world; if that weren't the case, why would I bother?
As such, it makes no sense to me for Google to introduce a tool that, by necessity, generates inaccurate and false views of the world. One of its selling points was viewing a place on the map as it looked in the past, but that's not what Nano Banana 2 was doing. Instead, the AI would pull from its training data to generate what it thought a historical representation of the area in question would be. If you asked it to show you what Paris looked like in 1726, it would likely overlay generic buildings and structures onto the map, rather than depict what Paris actually looked like three hundred years ago. What you're getting is the illusion of history; entertaining, perhaps, but certainly not educational—no better than scrolling through AI history slop on social media.
The same goes for the AI's "infographics." Companies like Google love to tout AI's potential as an educational tool, and on the surface, it seems like a great thing for Google Earth to generate learning graphics to teach you about a location on the map. But your generated graphic isn't vetted by anyone, and definitely not by an expert: The image that Nano Banana 2 spits out could be full of half-truths and hallucinations, and, what's worse, it won't offer any sources you can click through to check the AI's work. Casual users may retain the information they "learn" from these infographics, whether that data is correct or not.
At its worst, this feature was a disinformation machine. If you wanted to convince people that something bad happened in a particular part of the world, you could just type your request in, and Nano Banana 2 could spit out a realistic satellite image of your idea for easy sharing on social media. At a time when AI is spreading misinformation at a rapid rate, this only pours fuel onto an already raging fire. Companies like Google should be looking for ways to separate AI slop from real information, not helping it grow.
This isn't an argument against AI tools in their entirety, but it is an argument against embedding generative AI into an educational tool. What utility Google believes AI serves by taking truth and clouding it in falsehoods I do not know—other than giving it yet another example it can point to in order to tout its AI "success."
