TikTok's AI slop problem, by the numbers: 59% vs YouTube's 21%
A June 2026 study by Kapwing, also covered by Tubefilter and 9to5Mac, found that 59% of the first videos shown to a brand-new TikTok account qualify as AI slop, against 21% on YouTube Shorts using the identical method. The number is getting quoted a lot, usually with the wrong meaning attached. Here is what it actually measured.
The numbers, precisely
| Metric | TikTok | YouTube Shorts |
|---|---|---|
| AI slop rate, first 500 videos to a new account | 59% | 21% |
| AI slop rate in content aimed at kids | 57% | — |
| Total videos manually reviewed for the study | 10,742 across 20 categories | |
Kapwing defined AI slop as careless, low-quality video made with minimal human oversight, and classified each video by manual review against that definition rather than an automated detector. The data was collected in May 2026. Nearly three times as much of what a fresh TikTok account sees clears that bar compared to a fresh YouTube Shorts account.
What the number is not saying
The most common misreading turns "59% is slop" into "59% is AI-generated," which the study does not claim. It measured a quality judgment, careless and low-effort, not a production method, which is the exact distinction our honest slop self-audit makes about individual videos: the tool used is not the variable, the effort and research behind it is. A channel using AI narration and generated visuals with real research and a review pass is not what this statistic describes, however it was made. It is also worth being precise about scope: this measures what a brand-new account's algorithm serves by default, not the full catalog of either platform.
Why the gap exists
The study itself does not assign a cause, but the platforms differ in exactly the ways that would predict this outcome. YouTube gates ad revenue behind an originality review under the Partner Program, and its 2025 inauthentic-content policy specifically targets template-driven mass output, both covered elsewhere on this blog. TikTok's monetization has leaned harder on raw reach and volume. When a platform's incentive structure rewards volume over originality, more of what gets produced clears the bar for "careless," which is the finding here in one sentence.
What this means for where you post
It is one more data point in favor of long-form YouTube as a channel's primary home, the case already made on economic and AI-citation grounds in Shorts vs long-form. It is not a case for avoiding TikTok: the same cross-posting guide still applies, and a lower average bar on a platform is an argument for showing up there with something genuinely made, not for staying away. A channel that would pass the five-question slop audit stands out further on the platform where 59% of the feed would fail it.
Where Thothium fits
Thothium is built against exactly the pattern this data describes: scripts ground in real research before narration, every scene stays editable for a human review pass, and nothing ships as an unreviewed template with the subject swapped. It is in free alpha, and the form below gets you a key.
Frequently asked questions
What exactly did the Kapwing study measure?
Kapwing manually reviewed 10,742 videos across 20 popular categories, then separately tracked the first 500 videos shown to a freshly created account on each platform's default feed, TikTok's For You page and YouTube Shorts. 'AI slop' meant careless, low-quality video produced with minimal human oversight, judged by manual review against that definition, not an automated detector.
Does this mean 59% of TikTok is AI-generated?
No, and that is the most common misreading of the number. The study measured AI slop specifically, careless and low-effort, not AI-generated content broadly, and it measured what a brand-new account gets shown by the recommendation system, not the platform's full catalog. A channel using AI tools carefully is not what this statistic is describing.
Why is YouTube's rate so much lower?
The study does not establish a single cause, but the platforms differ in exactly the ways that would predict this gap: YouTube's Partner Program requires an originality review before a channel earns anything, and its 2025 inauthentic-content policy specifically targets template-driven mass output. TikTok's monetization has historically leaned more on raw reach. Incentive structures that reward volume tend to produce more of what gets flagged as slop.
Does this affect where I should post my faceless videos?
It is one more data point favoring YouTube as the primary platform for original, researched content, which lines up with the case made in our Shorts vs long-form and cross-platform distribution posts. It does not mean avoid TikTok; it means a channel built on real research and a review pass stands out further there precisely because the average bar is lower.
Last updated August 15, 2026. Figures come from Kapwing's June 2026 "TikTok AI Slop Report," based on manual review of data collected in May 2026, as also reported by Tubefilter and 9to5Mac. Platform-recommendation behavior changes constantly; treat this as a snapshot, not a permanent ranking.