Quick summary: New digital research has revealed that 59% of the videos recommended to new TikTok accounts consist of low-quality AI-generated content, with children under two experiencing the highest exposure at over 57%. Pediatric experts warn that this industrial-scale saturation of poorly automated media disrupts critical early brain development by establishing incorrect neural connections during formative growth stages. This rampant spread of unverified material highlights an urgent need for proactive public policy and enhanced healthcare practice guidance to support parents in managing screen time and protecting infant well-being.
Nearly 60% of videos served to new TikTok accounts are low-quality AI-generated content, according to new research that has highlighted the scale of so-called AI slop on the platform.
A report by video editing platform Kapwing found that 59% of the first 500 videos shown on a new TikTok account’s For You page were AI slop, three times the proportion found in a similar test on YouTube, where a fresh account encountered 104 such videos out of the first 500 Shorts, or 21%.
Kapwing defined AI slop as careless, low-quality content generated using automatic computer applications, often distributed to farm views or influence opinion.
The study also found that 57.4% of TikTok videos aimed at children fell into the same category. BBC technology correspondent Joe Tidy has observed that backlash comments against AI posts often receive more likes than the original content.
Categories most affected
Kapwing examined 10,742 videos across popular tags in 20 categories. The Kids category had the highest proportion of AI slop at 57.4%, followed by Science and Education at 35.0%, Health at 33.8%, and History at 33.5%. In the top 9 categories, more than 1 in 10 videos were AI slop. Fitness, Music, and Fashion categories had far lower rates, at 1.6%, 1.5%, and 1.3% respectively.
Scientific facts lend themselves to animation, but researchers have warned that hastily produced, sensationalist material can undermine educational value and compete with authoritative sources. Brendan Gillis warned that AI-generated history is only as reliable as its training data, potentially introducing inaccuracies and biases.
Concern for children
Of 2,000 videos analysed in the Kids category, 1,147 were AI slop. The hashtag #cartoonkids was 97% AI-generated, with only three human-made videos out of 100 sampled. Many other child-related tags showed high proportions, though #babytok had relatively less around 1 in 10. Examples include videos misusing Sesame Street characters for counting lessons with mismatched numbers and poor animation, with comment sections often featuring nonsensical text to boost visibility.
Dr Dana Suskind, professor of pediatrics at the University of Chicago, described the issue as toddler AI misinformation at an industrial scale. She told Mother Jones: “Every experience is building a million new neural connections. You will be unintentionally wiring the brain in incorrect ways.”
A survey from the Family Online Safety Institute found that only 51% of parents use parental controls on tablets and 47% on smartphones. According to the UK’s 1001 Critical Days Foundation, over 70% of babies and under-twos use screens, with 1 in 10 regularly falling asleep with one. The foundation notes the first 1,001 days from pregnancy to age two are critical for brain development.
Platform response
TikTok has previously acknowledged the role of AI-generated content, launching tools in 2025 to let users control the amount of AI-generated content in their feeds, stating that many people enjoy content made with AI tools, from digital art to science explainers. By November 2025, the platform had labelled 1.3 billion videos as AI-generated. It tailors feeds based on signals such as likes, follows, and scrolling behaviour, initially showing popular content appropriate for a broad audience.
In May, TikTok adjusted AI-generated summaries after errors, such as misidentifying dancer Charli D’Amelio, introduced options to reduce AI content, and announced a $2 million fund for AI literacy and safety education.
Concerns about AI-generated content extend beyond TikTok. Meta’s Mark Zuckerberg has described generative AI content as the third phase of social media, yet inaccuracies remain a persistent issue. Jonathan Laramy, creator of the AI history channel Chloe VS History, told Sky News: “Yes, there are people out there that obviously don’t care about history, because there are mistakes left, right, and centre.”
Experts emphasise that video remains a nuanced human form of communication, and outsourcing production without sufficient oversight risks lowering standards. Kapwing’s data is correct as of May 2026.