Kurzgesagt is a popular Youtube channel with almost 25 million subscribers. They publish beautifully animated and well-researched explainers of everything under the sun, including politics, science, and the sun itself. A recent topic? AI slop strangling the internet—and even three-billion-view entertainment empires like Kurzgesagt.
As per their research, more than half of the internet’s activity already goes back to bots. But they’re not just crawling, observing, absorbing our data like they used to. Since the advent of ChatGPT and co, the bots have started making things, and it’s beginning to show. “In 2025, there were already well over 1,200 confirmed AI news websites publishing massive amounts of AI-generated misinformation and false narratives,” Kurzgesagt explains. Not only is this making research more difficult—especially since AI lies so confidently in our faces all the time—but it might be the start of an irreversible decline in the glorious stock of human knowledge we’ve built up online thus far.
The team explains that while researching a topic for a video, they ran into the usual problems with AI. Some sources were inaccurate or outright false. Eventually, they even discovered some of the quoted sources were actually already written by AI! But while Kurzgesagt scrapped the project and started over, someone else didn’t, and voilà, a popular Youtube video citing all kinds of wrong ideas showed up on an AI-run channel a few weeks later. “This is where the death of the internet begins,” the team says.
Because guess what AI will now provide as a source for that topic? That AI-generated video! Once AI starts mostly referencing itself instead of proper, human-vetted information, the average quality of what you find online will fall off a cliff—and given the vast quantities of information generated at rapid speed, it’s unlikely to recover any time soon.
The Great Library of Alexandria was one of humanity’s first attempts at compiling its collective knowledge. Part of it burned down, and all of it was lost eventually. Now, we are “letting [AI] add new shelves to the library of human knowledge,” but since AI isn’t ready to do this responsibly, “the library of human knowledge is getting less and less reliable.”
We’ve already handed over a lot to Big Tech: our privacy, data, and attention. Now, it’s coming for our knowledge. Not by taking what we have but by choking it in a sea of artificial information garbage. It looks like the second time around, Alexandria will drown instead of burn—unless we deliberately choose human—human art, human ideas, human connection—first.
