TikTok's wider market list changes what its removal count means

The test

TikTok publishes a quarterly document it calls the Community Guidelines Enforcement Report, hosted on its transparency site as a living page rather than a series of dated PDFs. The edition current as of 16 September 2026 covers the January–March 2026 quarter, and the report states in its own methodology note that TikTok is 'significantly expanding the number of markets for which we provide enforcement data,' moving the threshold to any market with at least 50,000 removals in the quarter and raising the market count from roughly 50 to more than 130. The same note says TikTok is, from this report onward, folding videos removed by AI moderation into its automated-removals count for the first time.

What the evidence says

Both changes are TikTok's own methodology decisions, disclosed in the report rather than independently verified by an outside auditor. The report states that automated removals, AI included, now make up 'more than 96% of total removals.' A separate, undated statistics panel on TikTok's general transparency page states that 97% of violative video removals were automated, 99% were removed before a user reported them, and 96% of user-flagged videos were removed within two hours. Those headline percentages carry no quarter label on that page, so they should be read as TikTok's current, living claim rather than a fixed measurement tied to the same period as the quarterly report.

The sample and the variance

Because the market list nearly tripled between reporting periods and AI-removed content was added to a bucket that previously excluded it, a reader comparing this quarter's raw removal count, or its automated share, to an earlier quarter's would be comparing different populations, not just different results. The report breaks volume down by content type, policy category, market and language, but its scope is what TikTok itself chooses to disclose and label; it is not an independent audit of moderation accuracy, and it says nothing about violating content that is never found or reported.

What to try next

A reader tracking TikTok's enforcement trend over time should check each quarter's methodology note before comparing headline numbers, since TikTok itself flags when the counting rules change. This is an editorial habit, not a TikTok instruction: treat a jump in reported removals as a possible sign of a wider market list or a redefinition, not proof of a change in TikTok's underlying moderation behavior, until the methodology note says otherwise.

• Did the market list, the removal definition, or the content-type categories change between the two quarters being compared?
• Is a cited percentage from the dated quarterly report or from the undated statistics panel on TikTok's general transparency page?
• Does the automated-removal share include AI-detected takedowns for every period being compared, or only the newer one?

A rising removal count from a platform's own transparency report is evidence of what the platform disclosed, on the terms it set for counting. It is not, by itself, evidence that the platform grew safer or more dangerous.