TL;DR
Kapwing found that 59% of TikTok videos shown to new accounts had low AI, three times higher than YouTube, and that children’s content was the most hit category.
About six out of every ten videos TikTok serves to a new account are AI-generated garbage. This is the central finding A report published by video editing platform KapwingAnalyzing 10,742 TikTok videos across 20 popular categories and separately examining the first 500 videos displayed on the For You page of a newly created account.
Of those 500 videos, 294 were classified as AI downgrades, a Kapwing term defined as videos with obvious AI-generated visuals or low-quality compilations using AI-generated scripts and audio. 59 percent rate is approx That’s three times the rate found on YouTube in the same study, it dramatically degrades the default TikTok experience for anyone opening the app for the first time.
The numbers for children are even worse. Kapwing found that 57 percent of videos in TikTok’s Kids category were AI-friendly, the highest among the categories the researchers examined. Next comes science and education at 35 percent, health at about 34 percent, and history at about the same level.
💜 of EU technology
The latest rumblings from the EU tech scene, a story from our wise founder Boris and some questionable AI art. Free in your inbox every week. Register now!
At the other end of the spectrum, fitness, music and fashion content remained almost entirely human-generated, each below two percent.
One hashtag in particular shows the scale of the problem. Of the 100 videos checked under #CartoonKids, 97 were created by AI and only three were created by humans. Related hashtags were about equally bad: #cartoons and #babysong both reached 83 percent, and #children reached 79 percent.
The formula behind these videos is recognizable to anyone who comes across them. Familiar cartoon characters appear in strange scenarios, tutorials are full of factual errors, characters speak with synthetic voices, and animations shift and change in completely nonsensical ways. A counting lesson that gets the numbers wrong may seem absurd to an adult, but a preschooler has no context to record.
Dr. Dana Suskind, professor of pediatrics at the University of Chicago, describes this phenomenon as “industry-wide baby AI disinformation,According to the research report. The concern isn’t just that individual videos are bad, but that generative AI can create endless streams of them at a speed that no human creator can match.
The problem goes far beyond content aimed at children. Education, science, health, and history videos were among the most saturated categories with AI-generated content, which is especially damaging because these are the topics where accuracy is most important.
It’s easy to pass off a poorly crafted comedy skit. A history lesson filled with fictitious details or a health video that offers false advice is a different kind of failure.
TikTok’s recommendation engine is designed to adapt quickly, using signals like watch time, likes, follows and scrolling behavior to personalize what each user sees. But Kapwing’s research focuses on what happens before personalization begins, when the new account doesn’t provide behavioral data and the algorithm is essentially guesswork.
The result is that AI bias has become TikTok’s default first impression. For a platform that has built its growth on the strength of its recommendation algorithm, this is a significant challenge.
TikTok is not unaware of the problem. The company introduced controls in November 2025 that let users increase or decrease the amount of AI-generated content in their feeds, and it has invested in AI literacy initiatives. Kapwing argues that this passive control is insufficient, and data shows that the measures have not significantly reduced the amount of AI speed reaching new users.
The platform too faces increasing legal pressure over its handling of children’s content. Florida sued TikTok earlier this month under its children’s social media law, alleging the platform allowed minors to access the app and misled parents about the content available to them. The AI implications add another dimension to regulatory scrutiny: even when kids are on the platform legally, the content they encounter may be very low-quality and machine-generated.
The comparison with YouTube is instructive. Kapwing found that about 21 percent of YouTube Shorts recommended to a new account are AI-driven, less than half of TikTok’s rate. YouTube took a more aggressive enforcement approach in January 2026, suspending 16 channels with 35 million subscribers and nearly five billion lifetime views under its original content policy.
This push has drawn criticism for catching legitimate faceless creators in the crossfire, but the gap between platforms remains stark.
The broader pattern is consistent across social media. AI-generated content is flooding music streaming platforms and services like Deezer record more than 20,000 fully AI-generated tracks per day.
The incentive structure rewards volume over quality: if a creator or bot operator can produce dozens of videos in one shot, platforms are saturated with content that is technically watchable but offers little substance.
Limitations of the study should be noted. Kapwing is a video editing tool company that has a commercial interest in human-generated content and engages in the classification of what is considered “.AI bendingThe researchers created a master list of 20 popular TikTok categories, selected at least three popular tags for each, and reviewed the videos for obvious AI-generated visuals and scripts, a method that is transparent but subjective.
The study also paints a picture from May 2026, and TikTok’s algorithm and moderation policies may change. The platform findings were not publicly disputed.
Still, the scale of the data, over 10,000 videos across 20 categories and 500 video new account tests, make it the most comprehensive examination of AI content density on TikTok ever published. The findings of children’s content are difficult to dismiss regardless of the methodology: when 97 out of 100 videos in the child hashtag are machine-generated, the exact definition of “”slope” is less important than the fact that virtually nothing is human-made in that feed.
Social media became popular because it offered something distinctly human: creativity, identity, experience and connection. AI can imitate all of these with increasing skill, but imitation is not the same as originality.
With nearly six out of every ten videos seen by a new user being created by AI, the question is no longer whether TikTok has AI levels. The question is, has it become a defining feature of the platform?
