Overview
- The Kapwing analysis published Monday found that about 59–60% of videos shown to a newly created TikTok account were classified as AI-generated or AI-assisted after manually reviewing the first 500 For You posts.
- A parallel fresh-account test on YouTube Shorts showed roughly 21% AI-related videos, with Kapwing recording 104 AI-tagged items out of 500, suggesting TikTok’s exposure to synthetic media is about three times higher.
- The study expanded to more than 10,000 TikTok videos across 20 categories and found the highest concentrations in kids’ content, with hashtags such as #cartoonkids at 97% AI-generated and #babysong and #cartoons at about 83%.
- Kapwing reports that informational categories that use illustrations and voiceover—Science, Education, Health and History—have substantial AI presence of roughly one-third, while lifestyle and real-life categories like Fashion, Music and Fitness show single-digit rates.
- TikTok has previously labeled large numbers of AI videos and added user controls in November 2025, but the new findings highlight gaps in detection and moderation and renew calls for clearer labeling, better filters, and assessment of effects on young viewers.