- More than 20% of recommendations to new users on YouTube are low-quality AI-generated videos
- Spain leads in the number of subscribers to “AI slop” channels, exceeding 20 million.
- YouTube's recommendation and monetization model incentivizes the mass production of automated junk content
- This dynamic makes it difficult to distinguish real content from AI-generated content and degrades the informational value of the platform.
In recent months, many users have begun to notice that The YouTube feed, especially in Shorts, has become filled with strange, repetitive, and almost pointless videos.It's not just a feeling: several independent analyses indicate that a significant portion of what the platform recommends to new account holders is no longer produced by people, but by artificial intelligence systems that generate videos in sequence.
A recent report from the website Kapwing puts numbers to this phenomenon and names it “AI slop”: Massive amounts of AI-generated content, of little creative value and geared almost exclusively towards pleasing the algorithmThe study concludes that Over 20% of the first 500 videos suggested to a new user are already automated junk.And that countries like Spain have become one of the epicenters of this type of channel.
What is “AI slop” and why is it invading YouTube?
The term “slop” has become so popular that Merriam-Webster lists it as a synonym for low-quality digital content created in large quantities using AIIn the context of YouTube, these are videos produced almost without human supervision, based on clips generated by... models like Sora or VEO, synthetic voices and prefabricated scripts that are recycled over and over again.
These videos They do not aim to provide information, tell original stories, or experiment artistically.Its objective is much more prosaic: Maximize audience retention with exaggerated stimuli, frenetic rhythms, and narratives simplistic to the point of absurdity, a logic very similar to that of the “brainrot” content that swept across social media from 2025 onwards.
In the experiment that underpins the report, Kapwing created new profiles to simulate a "clean" algorithm and analyzed the first 500 YouTube recommendations. Around one in five videos was clearly AI-generatedwhile an additional third fell squarely into the category of trivial and overstimulated content designed to keep the user swiping without thinking too much.
The key is that Producing this material is extremely cheap and almost unlimited.Simply combine a more or less generic text, a video generator, a speech synthesis tool, and a few templates optimized for the algorithm. The result is thousands of clips that look very similar to each other that are uploaded automatically, trusting that the recommendation system will end up rewarding a few.
Spain tops the ranking of AI-generated waste
One of the most striking conclusions in the report is that Spain leads the world ranking of subscribers to channels classified as “AI slop” on YouTubeAdding together the audiences of these types of channels, the country would exceed 20 million subscriptions, above the United States (around 14 million) and Brazil (around 13 million).
This data was obtained by analyzing The 100 channels that appeared most frequently in trending topics in each countryA total of approximately 15.000 channels were analyzed. Of these, around 278 were categorized as "pure slop," meaning content produced almost entirely by AI and with a systematic pattern of repetition, exaggeration, and little informative or educational value.
For the Spanish-speaking market, the report identifies Several channels geared towards Spanish and Latin American audiences are among the most prominentSome present themselves as children's channels, others as light entertainment or supposed educational content, but they all share the same pattern: Short, highly addictive videos produced at top speed with AI.
A particularly representative example is that of story and animation channels inspired by popular universes such as Dragon Ball or other well-known franchises. These channels amass millions of followers and billions of views using recognizable AI-generated characters and styles, mixed with simplistic plots and an aesthetic that, if you scratch the surface, reveals the automatic stamp.
How are these trash videos made with AI?
The process for producing this type of content is becoming increasingly standardized. Many creators - or rather channel operators - start with texts in the public domainsuch as classic tales, fables, or legends. From there, they feed a video generation tool to obtain animated scenes, add automatic narration and background music, and in a matter of minutes they have a new clip ready to upload.
According to investigators, Generative video models allow you to create complete scenes in just a few clicks.Therefore, the bottleneck is no longer production, but the ability to upload and optimize videos. Many channels even automate this part, scheduling constant uploads and using titles and thumbnails designed exclusively to attract a large number of clicks.
The result is a A perfect storm of automation, algorithmic optimization, and degradation of educational and informational valueIn tests with new accounts, it is estimated that between 21% and 33% of the first 500 recommended videos fall under the "slop" label, either because they are generated directly by AI, or because they replicate its logic of saturation and constant stimulation.
All of this, moreover, is usually done without clear labels indicating that the content is AI-generatedThe warnings, when they exist, are discreet or simply nonexistent, so many users cannot distinguish whether there is a human creator behind it or an automated video production chain.
A lot of money is at stake and little interest in putting a stop to it.
Behind this avalanche of AI-generated garbage there is an obvious reason: The business is workingSome of the channels mentioned in the report, especially in markets with large audience volumes, can reach to generate several million euros a year in advertising revenueespecially if they combine direct YouTube monetization with brand deals and other business avenues.
Cases are cited such as certain animation channels for children or families with more than 2.000 billion accumulated views, whose Estimated revenue easily exceeds millions of dollars annuallyEven if the content is poor and repetitive, as long as it generates views and keeps the user on the platform, the business model works.
This explains why There doesn't seem to be any decisive move on the part of YouTube to contain the spread of this phenomenon.As long as the videos respect the basic community guidelines—do not incite hatred, do not include explicit violence, etc.—they can continue to be monetized and, in many cases, boosted by the recommendation system itself.
The platform thus faces a fundamental contradiction: Limiting the “AI slop” could mean giving up a significant portion of their revenue and overall viewing time.And, for the moment, the balance seems to be leaning more towards letting the automated waste market self-regulate than towards introducing strict filters that prioritize quality and transparency.
The problem isn't AI, it's the recommendation and business model
Most of the experts consulted agree that Artificial intelligence itself is not the enemy.The same tools that are used today to fill YouTube with empty videos could be used, in the hands of responsible creators, to experiment with new formats, translate content, or facilitate independent audiovisual production.
The focus, they point out, should be on the recommendation and monetization model of the platform itselfThe algorithm prioritizes quantity, frequency of publication, and the ability to hold attention over other factors such as originality, informational quality, or context. In that environment, The mass production of AI-generated videos fits like a glove..
In markets like Spain, where the same video in Spanish can go viral simultaneously in several countries, the effect is multipliedA single "AI slop" channel, well-tuned to the algorithm, can build gigantic audiences in Spain and Latin America simultaneously, further reinforcing the incentives to continue producing automated garbage.
If YouTube does not introduce fundamental changes in how it decides what to recommend and how it distributes monetization, The risk is that the platform will end up saturated with empty content.It will become increasingly difficult to find videos created by people with their own ideas, original approaches, or informative rigor, especially for newcomers who depend almost entirely on the system's initial suggestions.
From the user's perspective, this presents a scenario in which The ability to distinguish between genuine content and mass-produced AI-generated pieces becomes crucialThe overabundance of videos that appear human, but are not, can erode trust in what is seen and heard, and further complicate the already delicate information ecosystem on the internet.
Given the data that indicates Spain as a leader in subscriptions to “AI slop” channels, and the proportion of automated videos that slip into recommendations for new users, everything points to the fact that AI-generated garbage on YouTube has gone from being a rarity to a structural part of the platform's landscape.As long as the business model continues to reward quantity and retention over quality and transparency, and there are no clear measures to visibly identify what is created with AI, the challenge for those seeking reliable content – both in Spain and the rest of Europe – will be learning to navigate an environment increasingly crowded with videos that are shiny on the outside but hollow on the inside.
