Low-quality AI videos are taking up a striking share of TikTok’s For You feed, and the pattern is especially visible in content aimed at children. Kapwing’s latest research found that 59% of videos appearing on TikTok’s For You page were low-quality AI content, compared with 21% on YouTube.
The findings point to how recommendation systems can amplify automated videos once they start matching broad audience signals. They also raise a sharper concern around children’s categories, where machine-made clips appear to be spreading fastest.
Why the feed is filling up
Kapwing says TikTok recommends videos using user interests, watch activity, followed accounts, and interactions such as likes. New accounts are also shown popular content that the platform considers broadly relevant, including material influenced by location and language settings.
In tests with new accounts, researchers found that low-quality AI videos still surfaced frequently. The content appeared across multiple categories, including children, science and education, health, and history.
Children’s content stands out most
Among 10,742 videos analyzed across several popular TikTok tags, the children’s category had the highest concentration of low-quality AI content at 57.4%. That put it well ahead of the other categories studied.
Science and education followed at 35%, while health reached 33.8% and history came in at 33.5%. The data suggests that automated content is no longer limited to entertainment spaces and is moving into areas often treated as informative or educational.
Some categories still showed a stronger presence of human-made videos. Fitness, music, and fashion were among the areas where AI use was lower than in the rest of the sample.
A troubling concentration on child-focused hashtags
The sharpest warning sign appeared in the #cartoonkids hashtag. In that sample, only three out of 100 videos were made by humans, while 97% were low-quality AI content.
That scale of saturation is alarming because it places mass-produced visual material directly in front of young users. The concern is not only the volume, but also the possibility that children keep seeing repetitive content that lacks strong editorial quality.
Experts warn about development risks
Universities and child development experts are also paying attention to the pattern. Dana Suskind, a professor of pediatrics at the University of Chicago, said the trend could interfere with children’s development if it continues unchecked.
“I call this industrial-scale AI disinformation for children. It is a risk to brain development,” Suskind said in the research.
The warning underscores that the issue goes beyond feed quality. As TikTok’s algorithm continues to elevate popular material, low-quality AI videos in child-sensitive spaces are becoming a problem that is harder to ignore.
