Low-quality AI-generated “brainrot” content is flooding YouTube, and the algorithm seems perfectly fine with it.
If your Shorts feed feels louder, lazier, and increasingly synthetic, you’re not imagining it.
A new study from video-editing platform Kapwing found that more than 20% of YouTube Shorts are now AI slop, defined as careless, low-effort content generated with AI and designed purely to farm views, subscribers, and ad revenue. Even more striking, 33% of Shorts fall into the “brainrot” category: compulsive, nonsensical, low-quality videos optimized for endless scrolling rather than meaning.
In other words, over half of Shorts content now sits somewhere between automated filler and digital junk food.
Kapwing analyzed the top 100 trending YouTube channels in every country, identifying which relied heavily on AI-generated content. It then cross-referenced those channels’ views, subscriber counts, and estimated yearly revenue using social analytics tools.
To remove personalization bias, Kapwing also created a brand-new YouTube account and tracked the first 500 Shorts served to it, offering a rare, unfiltered look at what YouTube surfaces by default.
What a “clean” feed actually looks like
On a fresh account, YouTube served its first AI slop video after just 16 Shorts. From there, it became routine. Out of the first 500 videos:
- 21% (104 videos) were low-quality, AI-generated slop
- 33% (165 videos) qualified as brainrot
That means more than half of what new users see is algorithmically rewarded low-effort content. This isn’t edge-case behavior. This is the system working as designed.
The study also shows just how large, and profitable, this ecosystem has become. Spain leads in AI slop consumption, with over 20 million subscribers to trending AI-driven channels, Egypt ranks second and The United States comes in third.
The most popular AI channel identified, Bandar Apna Dost, has amassed 2.07 billion views and is estimated to generate $4.25 million per year.
This isn’t spam. It’s a scalable, low-cost content model being actively rewarded.
Why the algorithm embraces it
YouTube doesn’t promote content because it’s good. It promotes content because it performs. AI slop works because it’s: Cheap or free to produce, infinitely scalable, optimized for watch-time loops and “engaging enough” to keep users scrolling.
As long as people watch, even passively, the algorithm reinforces the behavior. For creators, there’s little incentive to slow down. For platforms, there’s little incentive to intervene.
Labeling won’t fix the problem
Some platforms are beginning to respond. Meta and Pinterest are experimenting with AI labels, and Google’s Gemini can now detect AI-generated video. But labeling doesn’t change incentives. As long as consumption remains high, AI slop will keep winning.
This isn’t just about bad content. It’s about what happens when volume beats value and effort stops mattering. AI isn’t killing creativity. The algorithm is simply telling us what it wants more of, and right now, it’s slop.