AI slop is taking over YouTube, and the algorithm is doing exactly what it was built to do

💡

Key Takeaways:
• One in five videos recommended to new YouTube accounts is now AI-generated junk, according to Kapwing research on 15,000 channels.
• The top 278 "slop" channels are pulling in an estimated $117 million a year while racking up 63 billion views.
• The problem isn't spread evenly. Spain leads the world with over 20 million subscribers to these channels—that is nearly half the country's population.

Open a fresh YouTube account today, and there's a one-in-five chance the platform will serve you what the internet has learned to call AI slop, which means low-quality, AI-generated videos designed purely to farm views and ad revenue.

That's not a glitch. It's the system doing exactly what it was told to do. Research from video editing platform Kapwing, published in November 2025, examined 15,000 of YouTube’s most popular channels and found 278 that produce nothing but AI-generated content. Together, those channels have amassed 63 billion views and an estimated $117 million in annual revenue.

When researchers created a brand-new account to see what a fresh user actually experiences, 104 of the first 500 videos recommended were AI-generated. Roughly a third fell into what the internet calls brainrot, content that teaches you nothing and entertains you even less, but somehow keeps you watching.

This creates a massive first-impression problem. For the next billion people coming online, most of them in emerging markets, YouTube will not feel like a place to discover human creativity. It'll likely feel like a feed of noise that was never touched by a human hand.

MORE INSIGHTS ON THIS TOPIC:

Where is AI-generated YouTube content being made, and who profits from it?

The way this content spreads tells a story about money, not just technology. While “AI slop” ends up on screens everywhere, the people making it and the people watching it are often in very different places.

Spain currently leads the world with more than 20 million subscribers to AI-generated channels. Egypt follows with nearly 18 million, then the United States and Brazil. The audience is global. The incentives are not.

Production tends to concentrate where the profit margins are highest. In countries where YouTube ad revenue pays far more than the average local wage, AI slop factories have become a modern gold rush. India’s top AI channel, Bandar Apna Dost, has surpassed 2.4 billion views. In a country where the average annual income is roughly $2,400, a channel like that could be generating more than $4 million a year.

That gap makes it nearly impossible for human creators to compete. If you're filming, editing, and trying to tell a real story, you cannot keep pace with scripts that generate hundreds of videos a day for pennies. The result is a flood of bizarre, high-contrast nonsense: monkeys acting like people, disaster clips that never happened, endless loops of robotic narration.

It makes no sense to a human. It makes perfect sense to an algorithm that only cares about keeping your eyes open.

Why does YouTube’s algorithm push AI slop to new users?

YouTube’s recommendation engine drives roughly 70% of what gets watched on the platform. But it depends heavily on one thing: your history. When a new user logs in without that history, the algorithm simply grabs whatever is “working” globally. Right now, that means AI-generated slop.

For a kid opening YouTube for the first time in Jakarta or Lagos, this sets a dangerous tone. If the first ten videos you see are synthetic junk, you don't learn how to look for quality. You learn to passively consume noise.

And there's little you can do to fight it. Mozilla’s research shows that tools like “Not Interested” barely work. You can reject one bad video and be served dozens more just like it hours later. As long as people keep clicking, the system believes it's doing its job.

Is the internet splitting into human content and machine-made noise?

This brings us to a strange turning point for the web. We're entering a world where “content” no longer automatically means something a person made.

What's emerging looks like two internets. In wealthier markets, where users and advertisers pay a premium, platforms may retreat toward safety. “Verified human” content could become a luxury product, something you pay for or access through gated communities just to escape the noise.

In emerging markets, though, where data is expensive and algorithmic feeds dominate, the open internet risks turning into a landfill. If you cannot afford premium filters or subscriptions, your feed becomes an endless stream of synthetic filler that exists only to show you ads.

As 2026 approaches, the most important question may not be how many views a video gets, but whether you can trust that a human made it at all. YouTube now has a decision to make. It can keep chasing the engagement metrics that built its empire and slowly turn into a library of digital waste. Or it can change the incentives and decide that human provenance actually matters.