Since we have models of image generation and videos with AI so cutting edge, we have begun to get used to seeing all types of content generated by artificial intelligence on social networks. It’s no longer just that we eat the occasional deepfake, it’s that much of the entertainment content that appears in our feed has been generated by AI.
An analysis of more than 10,000 videos reveals that TikTok, the most susceptible to what we know as ‘AI slop‘, shows this type of content at a rate three times higher than that of YouTube, and that children’s content is the most affected.
The problem has a name. As we have mentioned before, It’s called AI slop, and it’s Pretty cheesy, mass-produced content with artificial intelligence tools with the sole objective of accumulating views. In this content it is very common for us to see cartoon characters in absurd situations, videos that aim to give us educational lessons, or voices generated by AI over images that are deformed, among many others.
There is now data that measures the presence of this type of content on TikTok, with a rather uncomfortable precision.
What the study says. Kapwing, a San Francisco-based video editing company, has analyzed 10,742 videos spread across 20 TikTok categories and placed special emphasis on the first 500 videos that the platform showed to newly created accounts. 294 of those 500 videos, that is, 59%, were classified as AI slop.
The study states that the methodology was manual and without automation, and the data collected corresponds to May 2026. Kapwing defines this type of content as videos with obvious use of AI-generated images, or low-quality compilations with clearly artificial scripts and voice-overs.
In perspective. That rate is almost triple what Kapwing recorded on YouTube when performing the same test with a new account. On the Google-owned video platform, 21% of the first 500 videos in the Shorts feed were AI slops. YouTube has also taken more aggressive measures, last January canceling 16 channels with a combined 35 million subscribers for violating its artificially generated content policy.
On TikTok, on the other hand, the phenomenon seems to have established itself as the natural state of the algorithm before the user even begins to personalize their experience by scrolling.
Children, the most exposed. The category with the highest slop density on the entire platform is children’s content, with 57.4% of videos classified as AI-generated garbage. Within the #CartoonKids hashtag, 97 out of every 100 videos analyzed were artificial. Furthermore, as the study points out, in #babysong and #cartoons, the proportion was around 83%. In #forkids, 79%. In this content it is very common to see characters from well-known series getting into situations that make no sense, lessons to be learned that make mistakes, or animations that mutate in a strange way.
Exposure to AI from an early age. The Next Web medium share the words of Dr. Dana Suskind, a professor of pediatrics at the University of Chicago, who describes the phenomenon as “AI disinformation for young children on an industrial scale.” “Each experience builds a million new neural connections. Without meaning to, you will be wiring the brain incorrectly,” he warned.
The risk is not limited to meaningless content, since the categories of science, education, health and history also register high rates of slop, between 33% and 35%, according to the study.
The algorithm amplifies the problem. According to Kapwing’s own study, when a new account shows interest in AI content, the algorithm interprets that signal and ends up showing more. TikTok’s personalization mechanism, designed to fine-tune the feed based on user behavior, turns the initial slop into a starting point that feeds back.
As well as points out The study, by the time the platform activated a control in November 2025 that allows users to increase or reduce the amount of AI-generated content in their feed, had already labeled 1.3 billion videos as AIGC (AI-generated content).
What TikTok has done. In addition to feed control, the platform has allocated two million dollars to an educational fund to develop content on literacy and AI. But the study’s numbers suggest that these measures have not significantly curbed the volume of slop reaching new users. Furthermore, having an inexhaustible source of content, such as that generated by AI, contributes to the retention of the platform, so it is understandable that the platform does not have much interest in passing the broom.
Cover image | Kenneth Schipper and ROBIN WORRALL
GIPHY App Key not set. Please check settings