YouTube has a growing problem on its hands. Research shows that more than one in five videos recommended to new users are what experts call “AI slop” – low-quality, artificially generated content designed purely to rack up views and ad revenue.
A recent study by video-editing company Kapwing examined 15,000 of YouTube’s most popular channels worldwide and uncovered a sprawling network of AI-generated content that’s capturing billions of views. The findings paint a concerning picture of how artificial intelligence is reshaping what people watch online.
The numbers are staggering. Researchers identified 278 channels dedicated entirely to AI slop, collectively amassing over 63 billion views and 221 million subscribers.
These channels generate an estimated $117 million in annual revenue, all from content that requires minimal human creativity or effort to produce.
How Low-Quality Content is Hijacking YouTube?
To understand what new users experience, Kapwing created a fresh YouTube account and analyzed the first 500 videos recommended.
The results were eye-opening: 104 of those videos were AI slop, while a full third fell into the broader category of “brainrot” – low-quality content designed to monetize attention spans.
These AI channels aren’t confined to one region. They’re globally distributed and globally consumed. Spain has been particularly affected, with 20 million people nearly half the country’s population, following trending AI channels. Egypt has 18 million followers, the United States 14.5 million, and Brazil 13.5 million.
Some channels have achieved massive success. Bandar Apna Dost, based in India, leads the pack with 2.4 billion views.
The channel features bizarre adventures of an anthropomorphic monkey and a Hulk-like character who battles demons while flying on a helicopter made of tomatoes. Researchers estimate it could be earning up to $4.25 million annually.
Other successful channels target children. Pouty Frenchie, based in Singapore, has accumulated 2 billion views with stories of a French bulldog visiting candy forests and eating crystal sushi, often accompanied by children’s laughter. It reportedly makes nearly $4 million per year. Meanwhile, Cuentos Facinantes from the US has become the most-subscribed channel in the study with 6.65 million subscribers, using cartoon storylines aimed at young viewers.
Inside the Semi-Structured Industry of AI-Generated Content
The content takes disturbing turns, too. The AI World, based in Pakistan, produces AI-generated shorts depicting catastrophic flooding with titles like “Poor People, Poor Family” and “Flood Kitchen,” often set to relaxing rain sounds meant for sleep.
Behind this phenomenon is a semi-structured industry of creators who share strategies across platforms like Telegram, WhatsApp, and Discord. They exchange tips, sell courses, and identify what they call “niches, specific types of content that perform well algorithmically.
“There are these big swathes of people exchanging tips and ideas about how to make slop that will be engaging enough to earn money,” explained journalist Max Read, who has extensively covered AI slop. “One niche I noticed recently is AI videos of people’s pressure cookers exploding on the stove.”
Many creators come from middle-income countries with strong internet connectivity, where YouTube earnings can exceed local median wages. India, Ukraine, Kenya, Nigeria, Brazil, and Vietnam are hotspots for AI slop production.
The work isn’t without challenges. YouTube and Meta’s creator programs lack transparency about payments, and the ecosystem is full of scammers selling get-rich-quick courses that often earn more than the content creators themselves.
How Algorithms are Redefining Content Quality?
What matters most isn’t human creativity but understanding the algorithms that distribute content. These platforms function as massive testing machines, constantly evaluating what captures attention.
“The question is, how do you find the things that are doing well, and then how do you scale that? How do you make 10 of them?” Read said.
YouTube maintains that generative AI is simply a tool that can produce both high and low-quality content. A spokesperson stated the platform remains “focused on connecting our users with high-quality content, regardless of how it was made,” and removes content that violates community guidelines.
Yet the research suggests YouTube’s systems are actively recommending AI slop to new users, raising questions about whether the platform can effectively distinguish quality content from algorithmically optimized noise. As AI tools become more sophisticated and accessible, this challenge will only intensify.
