YouTube Deleted 16 Top AI Slop Channels With 4.7 Billion Views: What Happened and How to Stay Safe [2026]

YouTube AI slop purge: creator and cute robot beside colorful screens as 16 AI channels are removed

If your Shorts feed has started to feel like the same robot voice reading the same script over a recycled cartoon, you are not imagining it.

Sixteen of the biggest AI-driven channels on YouTube have now been deleted or stripped of every video. Together they had about 35 million subscribers and about 4.7 billion lifetime views.

That is a huge number. But the story is bigger than the numbers.

This post sorts the facts from the noise. You will see what the reports actually say, what YouTube has said on the record, and what it all means if you run a channel, work with AI tools, or just watch a lot of video.

Quick Facts at a Glance

What	Reported figure	Where it comes from
Channels removed or emptied	16 of Kapwing’s top 100 AI slop channels	Kapwing, Android Police
Fully deleted	11 channels	Kapwing update
Videos wiped, channel left up	5 channels	Kapwing update
Combined subscribers	About 35 million	Kapwing update
Combined lifetime views	About 4.7 billion	Kapwing update
Combined annual earnings	About $9.8 million (an estimate)	Kapwing, using Social Blade midpoints

What Actually Happened: The Purge by the Numbers
What Is “AI Slop” and How Was It Measured?
Why Did YouTube Act Now?
The Monetization Rules Creators Must Know
YouTube’s Automatic AI Labels Explained
Faceless Creators and False Positives
A Safety Checklist for Your Channel
What This Means for YouTube’s Future?
FAQ

1. What Actually Happened: The Purge by the Numbers

Let me start with where the story comes from, because that matters.

In late 2025, the video platform Kapwing published a study listing the 100 most-subscribed YouTube channels it classed as AI slop. A couple of months later, the tech site Android Police noticed that 16 of those 100 channels were gone.

So this was not a press release from YouTube. It was a list that someone checked again.

The 16 channels in plain numbers

Here is what Kapwing’s follow-up update reported for the 16 channels as a group:

11 were deleted completely
5 had every video removed but still exist as empty channels
Together they had about 35 million subscribers and over 4.7 billion lifetime views
Their combined estimated annual earnings were about $9.8 million, according to Kapwing’s estimate based on Social Blade data

That last point deserves care. The money figure is an estimate, not an audited number. Treat it as a rough size of the business, not a bank statement.

The channels named in reports

Three of the biggest names appear in most coverage:

CuentosFacianantes had about 5.95 million subscribers and published AI-made Dragon Ball style shorts
Imperio de Jesus had about 5.87 million subscribers
Super Cat League had about 4.21 million subscribers

Two of the top five channels on Kapwing’s list were among the 16. Reports also say the channels were based in several countries, including the United States, Spain and Indonesia.

What “removed” really means?

This is where many posts get sloppy.

“Removed” covers two different outcomes. Eleven channels vanished. Five were emptied out but left standing.

And one more thing: reports describe the removals as something people spotted by checking the list again. YouTube has not, in anything I found, published a channel by channel reason. So you should read the timing as informed context, not confirmed motive.

🔗 Related read: Not sure whether a video you watch is made this way? Here are the signs you are looking at AI slop and a quick way to check before you subscribe.

2. What Is “AI Slop” and How Was It Measured?

“AI slop” is not an official YouTube category. It is a slang term for mass-produced, low-effort content made mostly by generative tools, usually to collect views.

The reported traits are easy to recognize:

Synthetic narration with no human host
The same visuals reused across hundreds of videos
One formula repeated with tiny changes
Template-style production that needs almost no effort
Huge upload volume aimed at the recommendation system

What Kapwing actually measured?

Kapwing’s research is the backbone of this story, so here is what it did and did not say.

In its AI slop report, the company found that about 21% of the Shorts shown to a brand new user were AI slop. Other coverage cites a range of 21% to 33% when brainrot style content is included.

Notice the wording. That is what a new account saw in its feed. It is not the share of all Shorts on YouTube.

Kapwing also looked at 15,000 of the most popular channels worldwide and flagged 278 that posted nothing but AI slop.

Why the method matters?

Kapwing decides what counts as slop. YouTube did not publish that list. So when someone says “YouTube purged AI slop,” the careful version is this: YouTube removed channels that a third party had labelled as AI slop.

That is still a major event. It just needs the right label.

🔗 Related read: Want the Shorts side of the story? See what the data shows in YouTube Shorts and AI slop by the numbers and why the feed feels so repetitive.

3. Why Did YouTube Act Now?

There was no single trigger. Three pressures built up at the same time.

Pressure 1: A stated priority from the top

In his annual letter for 2026, YouTube CEO Neal Mohan listed “managing AI slop” as a priority. Coverage of the letter notes that it was one of the first times a major platform used the phrase so openly.

In short, YouTube told the public this was coming.

Pressure 2: Children and recommendation feeds

In February 2026, The New York Times published an investigation into AI-made videos aimed at toddlers and preschoolers.

On April 1, 2026, the advocacy group Fairplay sent an open letter to YouTube and Google signed by more than 200 organizations and experts. It asked YouTube to label all AI content and to ban AI-generated videos from YouTube Kids.

The letter also cites reporting that about 40% of the videos recommended after popular preschool shows contained AI content. YouTube’s spokesperson replied that AI content in YouTube Kids is limited to a small set of high-quality channels.

Pressure 3: Trust and advertiser confidence

Trade coverage of YouTube’s later monetization changes ties them to protecting advertiser confidence. That is a sensible reading, but it is analysis, not a quote from YouTube about these 16 channels.

So the honest summary is this. YouTube said slop was a priority, children’s advocates raised the alarm, and the commercial incentives all pointed the same way.

🔗 Related read: The child-safety side of this debate is covered in YouTube’s AI slop kids ban explained, including what parents can check today.

4. The Monetization Rules Creators Must Know

Here is the first thing to get straight. Losing ads and losing your channel are two different problems.

The 16 channels above were removed or emptied. The rules in this section are about something else: whether a channel can earn money in the YouTube Partner Program. For most creators, losing monetization is the bigger everyday risk, since TechRepublic notes that the update does not ban AI videos.

What changed on July 16, 2026?

On July 16, 2026, YouTube published clearer guidance on its “inauthentic content” policy. That policy is the newer name for the older “repetitious content” rules.

Is it a brand new rule? No. Tubefilter’s coverage describes it as a communications fix, with clearer language so creators can see why a channel gets flagged.

That matters because many posts online call it a “massive new policy.” The accurate version is smaller and more useful: the same rule, finally explained in plain categories.

The three categories that cannot earn ads

Category	What it means	Does AI have to be involved?
Generic or repetitive content	Videos made again and again from AI, CGI or templates with barely any change, plus tutorials that copy material already widely available	No, but AI makes it easy to do at scale
Unsatisfying or off-putting content	Videos built to grab clicks through manipulation, such as staged animal distress followed by a rescue	No. It can lose monetization with or without AI
AI personas on sensitive topics	An AI character presented as a human expert giving advice on health, legal issues, finances or politics	Yes

What YouTube says about AI itself?

YouTube’s head of trust and safety, Matt Halprin, was careful not to frame AI as the villain. His point, in his words, was that “AI can actually allow people to make a lot of videos,” and that the same tool can also make lots of very similar ones.

That is the real line. AI as a helper is fine. AI as a replacement for original thinking is where the trouble starts.

🔗 Related read: Not sure if your channel is already at risk? Check these signs a YouTube channel is being demonetized for AI content before you upload your next video.

The part nobody can measure

Here is the honest catch. Halprin did not define a threshold. Coverage of the briefing notes that channels with too much repetitive, low-effort or manipulative content can lose monetization, but nobody has said how much is too much.

So you cannot count your way to safety. You have to judge it the way a reviewer would: does each video add something a viewer cannot get from the next ten channels?

A simple self-test

Ask yourself these five questions before you publish:

Could someone swap my script into another channel and nobody would notice?
Does my video show a real point of view, a real test or a real opinion?
Would the video still make sense if the AI voice were replaced by a human one?
Am I giving health, money or legal advice through an AI character?
Is my main goal the viewer’s benefit, or just volume?

If you hesitate on two or more, that video deserves another pass.

🔗 Related read: Losing a channel is not only a policy story. See how one creator lost a long-running channel in YouTube’s AI killed a 13-year-old channel and what the mistake teaches.

5. YouTube’s Automatic AI Labels Explained

The second big change is about honesty, not money.

Since 2024, YouTube has labeled videos when creators say they used AI tools. In May 2026, it went a step further. According to YouTube’s announcement, the platform began using internal signals to spot significant photorealistic AI. If a creator does not disclose it, YouTube can now apply the label itself.

Where viewers see the label?

On long videos, the label sits directly below the player and above the description
On Shorts, it appears as an overlay on the video
This is now the single label format for photorealistic and meaningfully AI-made content

Lightly edited or clearly unrealistic AI content still gets its label only inside the expanded description, according to reporting on the rollout.

When the label cannot be removed?

YouTube says the label is permanent in two cases:

Videos made with YouTube’s own tools, such as Veo or Dream Screen
Files carrying C2PA metadata that marks them as fully AI-generated

C2PA is an industry standard that attaches a kind of origin record to media. If a tool stamps your file with it, YouTube can read that stamp.

What the label does not do?

This is the point many creators miss. YouTube states that a disclosure label alone does not change how a video is recommended or whether it can earn money.

So a label is not a penalty. A label plus repetitive, low-effort content is a different story, because the monetization rules above apply on their own.

Labels versus mistakes

Automatic systems make errors. A creator who edits in the open and uses AI only for cleanup can still be misread.

That is why records matter. Keep your drafts, scripts, raw footage and editing notes. If a label or a monetization decision looks wrong, you want proof of the human work behind the video.

🔗 Related read: Mistakes by automated systems are a real risk. Read how Google’s AI deleted YouTube channels and what creators did to respond.

Quick comparison: label, demonetization and removal

	AI label	Loss of monetization	Channel removal
What happens	A transparency tag appears	The channel cannot earn ads	The channel is deleted or emptied
Triggered by	Significant photorealistic AI, disclosed or detected	Inauthentic content categories	Reported cases involved mass-produced AI slop
Hurts reach by itself?	YouTube says no	Not directly	Yes, the channel is gone
Can you appeal?	You can request changes if it is wrong, except permanent cases	Yes, through YouTube Studio	Depends on the case

6. Faceless Creators and the Problem of False Positives

This is the section most news posts skip, and the one many small creators care about most.

Faceless channels have existed on YouTube for years. Think voiceover explainers, ambient videos, animation and documentary-style channels where the host never appears on camera. Plenty of them are fully human work.

What the policy actually targets?

Read the three categories again. None of them mention faces. The target is repetition, manipulation and fake expertise.

The Desk’s coverage of the briefing describes the goal as separating creators who use AI to improve original work from operators who mass-produce similar videos for ad revenue. A faceless channel with original scripts and research sits on the safe side of that line.

I looked for official data on how often automated systems wrongly flag legitimate channels. I could not find any. You will see creators on social media describing unfair flags, but those are personal reports, not published statistics.

So the fair summary is this: false positives are a real worry, and nobody outside YouTube can say how common they are.

Where faceless channels can look risky?

Faceless does not equal risky. Certain habits do:

A fully synthetic voice reading text pulled from other pages
Stock or recycled visuals reused from video to video
Identical intro, structure and pacing every single time
Dozens of uploads a week with no clear editing effort
An AI character posing as a doctor, lawyer or financial expert

If three or more of those describe your channel, a reviewer could reasonably see mass production, even if you put hours into each script.

🔗 Related read: Thumbnails can also signal sameness to viewers and systems. See why AI-generated thumbnails can trigger an algorithm penalty and how to avoid the trap.

How to show real human work?

You do not have to show your face. You can show your fingerprints instead.

Keep script drafts, research notes and source lists
Save raw clips, project files and version history
Add original commentary, testing, screenshots or data to every video
Use your own voice, or a consistent narrator with a clear editorial style
Build a community page or newsletter so viewers know who runs the channel

None of this guarantees a good outcome. But if you ever need to appeal a decision, a folder of proof beats a vague explanation.

7. A Safety Checklist for Your Channel

Here is a practical checklist, split by how you use AI. Print it, paste it into your notes, or run through it before each upload.

If you use AI as a helper

Do this	Avoid this
Use AI for editing, captions, research support and drafts	Publishing raw AI output without reviewing it
Add your own opinion, experience or testing to every video	Using one template with only the topic swapped
Disclose AI use honestly through YouTube Studio	Hiding realistic synthetic footage, since automatic labels can apply anyway
Vary your formats, structure and visuals	Reusing the same imagery across hundreds of uploads
Put a real, visible human expert behind health, legal or money advice, as outlined in a July 2026 breakdown of the update	Using an AI avatar as the face of sensitive advice

If you run a faceless channel

Document your process from script to final export
Add an original element to each video, such as a test, an interview, a data pull or a personal take
Slow down your upload pace if quality is slipping
Keep the channel’s identity consistent: name, voice, style and “about” page
Check YouTube Studio notices regularly so policy emails never sit unread

If you make content for children

This is the most scrutinized area right now. The AP’s report on the advocacy letter notes that YouTube said it is working on AI labels for YouTube Kids. Heavy use of AI in child-directed videos invites attention.

Keep the content accurate, calm and clearly educational. Avoid fast, repetitive, hypnotic loops built only to hold attention.

🔗 Related read: Parents and creators alike should read how AI brain rot videos target children and what warning signs to watch for.

What to do if your channel gets flagged?

Read the notice carefully and note the exact policy named
Compare the flagged videos against the three categories in Section 4
Gather your proof of human work before you appeal
Fix the pattern, not only the one video, because reviewers look at the channel as a whole
Appeal through YouTube Studio and be specific, factual and calm

8. What This Means for YouTube’s Future?

Here is the tension at the center of the story.

YouTube is clearing out mass-produced AI content. At the same time, it is building AI into its own tools. Veo and Dream Screen are YouTube’s own creation tools, and videos made with them carry a permanent AI label. YouTube has also widened its likeness detection system so any adult can scan the platform for videos that may contain their face.

So the platform is not anti-AI. It is against AI used as a volume machine.

Why this matters beyond YouTube?

Think about the money. Mass-produced content only works when the cost of making a video drops below the ad revenue it earns. Once platforms limit the earning side, that math breaks.

🔗 Related read: For the bigger economic picture, see how the AI slop economy works and who profits from it.

What to expect next?

These are reasonable expectations, not promises:

Platforms will keep tightening monetization instead of banning AI outright
Labels and detection will get stricter, so honest disclosure becomes the safer habit
Original perspective will matter more than production volume
Child-directed content will face the most pressure and the most scrutiny

The numbers that matter

Item	Reported figure
Channels removed or emptied	16 of Kapwing’s top 100 AI slop channels
Subscribers across those channels	About 35 million
Lifetime views	About 4.7 billion
Estimated yearly earnings	About $9.8 million (Kapwing estimate)
AI slop in a new user’s Shorts feed	About 21%, per Kapwing
Monetization rule clarification	July 16, 2026
Automatic AI labels began rolling out	May 2026

Frequently Asked Questions

Did YouTube ban AI-generated videos?

No. YouTube has said AI is allowed as a tool. The monetization limits target repetitive, manipulative or fake-expert content, with or without AI.

Can I still earn money if I use AI in my videos?

Yes, if your videos add original value and avoid the three inauthentic content categories. YouTube has not published a numeric threshold, so quality and originality are your best protection.

Will an AI label hurt my reach or earnings?

YouTube states that a label alone does not change recommendations or monetization eligibility. Repetitive, low-effort content can still cause problems on its own.

Are faceless channels in danger?

Not by definition. The published categories focus on repetition, manipulation and AI personas giving sensitive advice, not on whether a face appears on screen.

Who decided these 16 channels were “AI slop”?

Kapwing created the original list of 100 channels. YouTube has not publicly published its own list or a reason for each removal in the reports I reviewed.

The Bottom Line

YouTube did not declare war on AI. It declared war on sameness.

The 16 channels in this story had huge audiences, but according to the reports, they relied on mass-produced content built for the recommendation system. The rules coming out of 2026 make that model harder to sustain.

For creators, the lesson is simple. Use AI to work faster, not to think less. Add something that only you can add.

Original voice, honest disclosure and real effort are the strongest protection you have.