Table of Contents
The $117 Million Question: What's Actually Happening on YouTube Shorts?
The Kapwing Study: How Researchers Tested the Algorithm
The Countries Consuming the Most AI Slop
The AI Slop Economy: How Channels Are Making Millions
YouTube's Response: New Policies, Labels, and Crackdowns
Why the Algorithm Keeps Feeding You AI Slop?
What This Means for Real Creators?
The Bottom Line: Is Half of YouTube Shorts Really AI Slop?
You've probably scrolled through YouTube Shorts and felt it. That strange, uncanny valley feeling. A monkey in a dramatic situation that looks almost real but something's off. A baby trapped in space. A French bulldog in a candy forest. Endless variations of the same template, over and over again.
You're not imagining things. And the numbers are worse than you think.
A bombshell study from video-editing platform Kapwing dropped in late 2025, and it sent shock-waves through the creator economy. The finding? More than one in every two YouTube Shorts recommended to new users is either AI-generated slop or brainrot.
That's right. Fifty-four percent of what the algorithm serves fresh eyes is low-quality, mass-produced, algorithm-baiting content.
And here's the kicker: the channels pumping out this content are collectively raking in an estimated $117 million per year.
Let's break down what the data actually shows, where the money's going, and what YouTube is finally doing about it.
The $117 Million Question: What's Actually Happening on YouTube Shorts?
Kapwing, a video-editing platform you've probably used to trim a clip or add captions, decided to do something simple: create a brand-new YouTube account with zero watch history and see what the algorithm recommended.
No personalization. No browsing history. Just a blank slate.
What they found was staggering.
The researchers analyzed the first 500 YouTube Shorts served to this untouched account. Among those 500 videos:
104 videos (21%) were classified as AI slop — low-quality, mass-produced AI-generated content designed purely to farm views and subscriptions
165 videos (33%) fell into the brainrot category — nonsensical, low-quality content that creates the effect of "corroding the viewer's mental or intellectual state"
Combined, that's 269 out of 500 videos — or 54% of the entire sample.
Let that sink in. More than half of what a new user sees on YouTube Shorts is algorithm-optimized garbage.
What the Numbers Actually Say: 21% AI Slop, 33% Brainrot
The study didn't just throw around vague terms. Kapwing defined its categories with precision.
AI slop refers to content generated using artificial intelligence tools with minimal human effort or creative input. These videos are:
Mass-produced at scale
Optimized purely for algorithmic engagement
Lacking in educational or entertainment value
Often featuring repetitive templates, AI-generated voices, or synthetic visuals
Brainrot is a broader category that often overlaps with AI slop. It refers to:
Compulsive, repetitive content
Nonsensical or low-quality production
Content that creates the effect of "corroding" the viewer's mental state
The study found that 33% of recommended Shorts fell into the brainrot category — meaning even content that wasn't technically AI-generated was still low-quality, engagement-farming drivel.
Kapwing researchers created a fresh YouTube account and reviewed the first 500 Shorts shown in the feed. The results were consistent across multiple tests, suggesting this isn't a fluke — it's a systemic problem with how the Shorts algorithm operates.
The Countries Consuming the Most AI Slop
The AI slop phenomenon isn't evenly distributed across the globe. Some countries are consuming — and creating — far more of this content than others.
Spain leads the world in subscribers to trending AI slop channels, with a combined 20.22 million subscribers. What's remarkable is that Spain achieved this with relatively fewer AI slop channels among its top 100 trending channels, suggesting concentrated consumption.
South Korea dominates in total views. AI slop channels there have generated over 8.45 billion views — far exceeding totals seen in most other countries. A single channel, Three Minutes Wisdom, accounts for nearly a quarter of the country's total AI slop views, with 2.02 billion views.
The United States ranks third globally by subscriber count, with approximately 14.47 million subscribers across trending AI slop channels. The U.S. has nine AI slop channels among its top 100 most popular channels.
Other notable markets include:
Brazil and the Philippines: also featuring prominently in global rankings
The AI Slop Economy: How Channels Are Making Millions
Here's where the story gets truly wild. These AI slop channels aren't just filling up feeds — they're printing money.
Kapwing's analysis identified 278 channels producing nothing but AI slop content. Combined, these channels had:
Let's look at some of the biggest players.
Bandar Apna Dost — an Indian channel featuring an AI-generated monkey placed in emotional or dramatic situations — is the most-viewed AI slop channel globally. With over 2.4 billion views, it's estimated to earn approximately $4.25 million per year. The channel uploads hundreds of near-identical videos with minimal variation.
Three Minutes Wisdom — a South Korean channel — has accumulated over 2.02 billion views and is estimated to earn over $4 million annually from advertisements. That's a single channel, producing low-effort AI content, making more money than many professionally produced media outlets.
Pouty Frenchie — featuring an AI-animated French bulldog in candy forests set to children's laughter — pulls nearly $4 million per year.
Cuentos Facinantes — a U.S.-based Spanish-language channel — has the highest subscriber count among AI slop channels at 5.95 million subscribers.
The Guardian's analysis of Playboard data confirmed that nearly 10% of YouTube's 100 fastest-growing channels worldwide published exclusively AI-generated content.
The economics are simple: AI tools make content creation cheap, algorithms reward volume, and ad revenue flows to whoever can produce the most watchable (or at least swipe-able) content.
YouTube's Response: New Policies, Labels, and Crackdowns
YouTube hasn't been sitting idly by. The platform has rolled out a series of policy changes aimed at curbing the AI slop epidemic.
The Monetization Crackdown
Starting July 15, 2025, YouTube renamed its "repetitious content" policy to "inauthentic content" and provided clearer definitions of what won't be monetized.
Content made primarily using:
AI-generated voices without genuine human commentary
Synthetic avatars or automated narration
Repetitive templates with minimal variation
...will be demonetized.
YouTube made clear that AI-assisted content is acceptable, but only if it includes considerable human input and genuine value.
Prominent AI Labels
In May 2026, YouTube announced that AI disclosure labels would be moved to much more prominent positions.
Traditional YouTube videos now show an AI label right below the video player
YouTube Shorts display an AI label as an overlay at the bottom of the video
YouTube also began automatically scanning for signs of AI usage and labeling videos even if creators didn't disclose them. This applies to:
Content created using Google's own AI tools (like Veo and Dream Screen)
YouTube CEO Neal Mohan used the term "AI slop" himself in his January 2026 annual letter, pledging to build on YouTube's spam and clickbait-detection systems to combat it.
The Problem for Faceless Creators
There's a catch. YouTube's crackdown on AI slop is catching legitimate creators in the crossfire.
Faceless channels — where no human host appears on screen — have existed on YouTube for years. Many are run by solo creators who prefer anonymity, producing voiceover-driven explainers, ambient videos, or niche educational content.
The problem is that AI text-to-video tools made it easy to flood the platform with faceless content at scale, and YouTube's response has been to tune its algorithm to favor videos with real human faces on camera. That distinction doesn't separate AI-generated from human-made — it separates on-camera creators from off-camera ones.
Why the Algorithm Keeps Feeding You AI Slop?
You might be wondering: if AI slop is so low-quality, why does the algorithm keep serving it?
The answer is simple: the algorithm optimizes for engagement, not quality.
YouTube Shorts operates as a swipe-based feed where videos auto-play without requiring a click. The algorithm optimizes for immediate retention — specifically, whether viewers watch past the first two to three seconds or swipe away.
AI tools excel at generating attention-grabbing visual hooks that stop the thumb for those opening seconds. The content doesn't need to deliver value. It just needs to prevent a swipe for 15 seconds.
Kapwing's research showed that in the scenario with a totally new account, it took just 16 Shorts before YouTube served an AI slop video. From there, it was a steady stream.
Short-form video platforms reward:
...which makes the format perfectly suited to AI-generated content that creators can produce and publish rapidly with minimal variation.
What This Means for Real Creators?
If you're a legitimate creator producing original content, the AI slop epidemic creates a challenging environment.
Your content is competing with algorithm-bait. The platform's recommendation systems are optimized for engagement metrics that AI slop channels have learned to game.
Discovery is harder. When 54% of what new users see is low-quality content, it's harder for quality creators to break through.
Monetization is changing. YouTube's new policies mean that AI-assisted content with human value may still be monetized, but purely automated content won't be.
The silver lining? YouTube is finally taking the problem seriously. The platform's moves to:
Demonetize repetitive, inauthentic content
...are steps in the right direction.
The Bottom Line: Is Half of YouTube Shorts Really AI Slop?
Let's answer the question directly.
Yes — 54% of YouTube Shorts recommended to new users is either AI slop or brainrot.
That breaks down to:
21% pure AI slop
33% brainrot
Combined, that's more than one in two videos.
The data comes from Kapwing's rigorous study of 500 Shorts served to a fresh account — a methodology that eliminates personalization bias and gives us a clean look at what the algorithm serves by default.
The AI slop economy is real, it's massive, and it's generating $117 million annually for a relatively small number of channels.
But there's hope. YouTube's new policies, prominent AI labels, and automated detection systems are starting to push back.
The question isn't whether AI slop exists — the data proves it does. The question is whether platforms like YouTube can balance the economics of short-form video with the need for quality, trustworthy content.
And that's a question only the next few years will answer.
This article was written based on data from Kapwing's AI Slop Report (November last), The Guardian's analysis, Mashable's coverage, and Search Engine Journal's breakdown of YouTube's policy responses. All data reflects findings as of the dates cited.