Kurzgesagt says its video on superpredators became the channel’s worst-performing upload since 2013, even though the audience data looked strong.
Click-through was above the channel’s average. So was watch time. Viewer response was positive. The video simply wasn’t getting recommended at the level the channel normally sees.
Kurzgesagt contacted YouTube and says the company confirmed the cause: automated detection tools had classified the channel’s hand-animated work as low-quality AI content. That classification appears to have throttled the video’s reach.
The channel has pulled the video and plans to make changes before reuploading.
Other channels have similar issues
Analytics firm vidIQ pointed out the part that should worry everyone else. Kurzgesagt has 24 million subscribers and a direct line to YouTube. A channel with 5,000 subscribers watching views collapse has no way to find out whether an automated system is responsible, and no realistic path to asking.
YouTube has been ramping up its AI-quality enforcement. In March, the platform started surveying some viewers on whether a video they were watching felt like AI slop, gathering signal on suspected synthetic content. It has also taken action against channels pumping out repetitive AI-generated uploads, while still permitting AI tools as part of a normal production workflow.
The Kurzgesagt case shows where that gets hard. Kurzgesagt animates in a flat, clean vector style with a calm narrator over explanatory graphics, which is exactly the aesthetic AI slop channels imitate because it’s cheap to fake. A detector trained to catch the imitation is going to catch the original that inspired it.
That’s not a fixable problem so much as a permanent tradeoff. Every threshold YouTube picks either lets spam through or catches legitimate creators, and the creators who get caught mostly won’t know it happened.
Image courtesy: Kurzgesagt