AI slop: why your feed filled with AI videos

You open Reels to watch a video and, three scrolls later, you're on the fourth chapter of a soap opera where a banana discovers that the passion fruit cheated on her with the orange. Nobody filmed that. An artificial intelligence generated each scene, and someone posted dozens of episodes in the same week.

This has a name: AI slop, the AI video made in series, fast and cheap, just to grab your attention. Here you'll understand where the word came from, why this type of video took over the feed, where Brazil fits in, and how to see less of it.

Quick summary: the answer in 30 seconds. AI slop is low-quality AI content mass-produced to generate views. It filled the feed because making video became almost free, the algorithm rewards what holds the eye, and these channels make money. Brazil is the fourth country in subscribers to these channels. You can reduce what appears to you in each app's settings.

Index

• What is AI slop, anyway?
• Why did the feed fill with AI video?
• AI fruit and AI soap operas: slop with a Brazilian accent
• Is every AI video slop?
• What's the real problem?
• What are the platforms doing?
• How to identify AI video and clean your feed
• What this changes for you
• Frequently asked questions

What is AI slop, anyway?

AI slop is content generated by artificial intelligence with little or no care, published in large quantities to earn clicks and money. It applies to text, image and audio, but what exploded was the short video. The word slop means swill: the leftover food given to pigs.

According to linguist Adam Aleksic, interviewed by MIT Technology Review, this use of "slop" was born on the 4chan forum in the early 2010s and became a label for anything mass-produced and without quality.

Alongside it appears brainrot. Kapwing, a video editing company that made one of the largest surveys on the subject, defines brainrot as the senseless video, often made with AI, that you watch non-stop without knowing why.

In one sentence: slop is the internet's swill, and the question is why there's so much of it.

Why did the feed fill with AI video?

Three things happened at the same time: making an AI video became almost free, social network algorithms reward any video that holds your attention, and these videos make money. When producing costs little and the return is high, those who want profit post non-stop, and the feed fills up.

Making became almost free

Until 2023, video generators made a few seconds of blurry image, with melting hands. In two years, tools like Sora, from OpenAI, and Veo, from Google, began generating realistic scenes, with sound, from a prompt (the text request you write for the AI). A soap opera required actors, camera and scenery. Now it requires a sentence.

Companies turned this into social networks. In September 2025, Meta launched Vibes, a feed only of short videos generated by AI inside the Meta AI app.

The algorithm rewards what holds the eye

The algorithm (the system that chooses what appears in your feed, like a waiter who brings back the dish you ate to the end) doesn't know if a video was made by a person or by an AI.

It measures how long you stayed, if you watched to the end, if you came back. Slop is tailor-made for this, with an absurd scene right in the first second. And, since generating is cheap, a video that works becomes a hundred variations the next day.

Kapwing created a new account on YouTube and noted the first 500 Shorts (YouTube's short, vertical videos). Result: 104 of them, or 21%, were generated by AI. In a separate count, 33% were brainrot. An account that hadn't yet taught the algorithm anything was already receiving one in five videos made by machine.

And it makes money

Kapwing's report, with data from October 2025, points to India's Bandar Apna Dost as the world's most-watched slop channel: 2.07 billion views of videos of a realistic monkey in human situations, with estimated revenue of $4.25 million per year.

AI fruit and AI soap operas: slop with a Brazilian accent

Brazil is one of the world's largest audiences for AI slop. In Kapwing's survey, slop channels among the country's most popular totaled 12.56 million subscribers, the fourth largest number among the countries analyzed, just behind the United States (14.47 million). Spain leads, with 20.22 million.

And Brazilians gave slop a format of their own: the soap opera. Searches for "frutas de ia" (AI fruit) on YouTube grew 4,700% in the last year, and "novela de ia" (AI soap opera) rose 550%, according to Google Trends (checked on 06/11/2026; the numbers are relative, not search counts).

It makes sense. The soap opera has a formula every Brazilian recognizes: betrayal, villain, twist at the end of the chapter. AI keeps the formula and swaps the actors for fruit, which don't charge fees. The absurdity grabs in the first second, and the next chapter's hook makes you follow the profile.

Is every AI video slop?

No. Slop is the video made without intention, in series, just to earn. There are creators who use the same tools with their own idea, defined aesthetic and hours of work per video. The difference isn't in the tool, it's in who uses it and for what, like the same pan serves the chef and the diner.

MIT Technology Review interviewed some of them. Artist Wenhui Lim, from Singapore, created the Niceaunties profile, with Asian ladies in fantasy worlds, and says a one-minute video can take hours or days.

The useful question, therefore, isn't "was it made with AI?", but "did anyone care about this?".

What's the real problem?

The problem isn't the silly video itself, it's the scale. When up to one in five Shorts on a new account is made by machine, children receive content that nobody reviewed, it becomes harder to trust what you see, and real creators lose space to those who post dozens of videos per day.

Children. Animated video is the easiest to generate. By YouTube's own policy, the AI notice appears on top of the video itself when the content is realistic; in animated video, it stays in the expanded description, which a child is unlikely to open. The Verge showed how this throws oversight onto parents. Our guide on the dangers of artificial intelligence for children helps.

Trust. The same tool that generates a talking fruit generates a real person saying what they never said. Seeing something many times makes the brain find it familiar, even if false. It's the gateway to the scams we explain in AI scam: cloned voice and deepfake.

Creators. Those who take a week to record a video compete for the same feed with those who post fifty per day.

What are the platforms doing?

Networks began to react, but without banning AI. The chosen path was to cut the money from those who mass-produce, label generated content, and give the user a button to see less of it. YouTube, TikTok and Pinterest have already announced official measures.

The effect already shows. The two slop channels with the most subscribers according to Kapwing, Cuentos Facinantes (5.95 million) and Imperio de Jesus (5.87 million), went offline, as reported by The Verge in January. Both addresses remained offline when we checked, on 07/11/2026.

There's a contradiction: the same companies sell the tools that produce slop. Perhaps that's why YouTube says what matters isn't the tool, but the original effort, according to an executive interviewed by Engadget.

How to identify AI video and clean your feed

No trick is foolproof, but some signs appear frequently, and each app has a setting that reduces what reaches you. The most efficient is to combine both: distrust videos that look factory-made and tell the algorithm you don't want to see them.

Signs the video may be AI:

1. It's very short. The longer, the more chance the AI will err.
2. It imitates security camera or old phone. Grainy, distant image hides defects.
3. The physics doesn't add up. Shadow that doesn't match the light, wrong reflection, object that changes shape.
4. People too perfect. Impeccable people where nobody would be like that, like a made-up person in the middle of a flood.
5. The profile has dozens of identical videos. Same scenery and same characters, published on the same day.
6. It has YouTube's notice. Look for the label of altered or synthetic content (made or changed by AI), on the video or in the description.

For text, the signs are different, and are in the guide on how to recognize text written by AI.

Settings that help:

• YouTube: in the video's three dots, use "Not interested" or "Don't recommend channel".
• TikTok: look for the AI content control in "Manage topics," being rolled out gradually. While it doesn't arrive, hold your finger on the video and tap "Not interested".
• Pinterest: in a Pin's three dots, ask to see less AI content, or adjust in "Refine your recommendations".
• Instagram and Facebook: in suggested videos, tap the three dots and choose "Not interested".

In one sentence: skipping quickly and marking "not interested" is worth more than watching with anger.

What this changes for you

AI slop won't disappear, because it still makes money. What changes is your role: the feed responds to what you do.

To go further, the artificial intelligence section gathers Kalkazu.org's guides.

Frequently asked questions

What is AI slop?

It's low-quality artificial intelligence content, published en masse to gain views and money. Slop means swill, the leftover food given to pigs. Today, the term refers mainly to short videos on social networks.

Why are there so many AI fruit videos?

Because it's cheap to make, grabs attention in the first second, and works in series, like a soap opera.

Can you remove AI video from the feed?

Not completely, but you can reduce it. TikTok and Pinterest have specific controls, and in all apps marking "Not interested" teaches the algorithm.

To close

The feed filled with AI video because someone discovered you can manufacture attention almost for free. But the attention remains yours: each video you skip teaches the algorithm to show something else.