TikTok’s sixth Digital Services Act transparency report, described on the company’s newsroom as covering the second half of 2025, is the official EU-period dataset this briefing uses. The newsroom page loaded for this item states that 178 million people in the EU were coming to TikTok every month at the time of that report. Between July and December 2025, TikTok said it removed around 112 million pieces of content that violated its Terms of Service, Community Guidelines, Advertising Policies, and TikTok Shop Policies. The total, it said, includes videos, LIVE streams, ads, product listings, and — for the first time in this series — comments.
Automation is the other headline TikTok chose. Automated systems, it wrote, actioned 93.8% of all violating content without human review, and 97.6% of those automated enforcement decisions were confirmed as correct in TikTok’s own precision testing. Those two percentages are not the same statistic: one is the share of actions taken without a human, the other is TikTok’s measured accuracy of those actions. Neither is an EU regulator’s audit. The newsroom post also says TikTok has published voluntary transparency reports since 2019 and that the DSA reports sit on top of that practice.
TikTok’s Transparency Center index, which loaded as a second official source, lists the families of reports the company maintains: Community Guidelines Enforcement, covert influence operations, child sexual exploitation, government removal requests, intellectual-property notices, information requests, DSA, Shop safety, the Code of Practice on Disinformation, EU monthly active recipients, the Digital Markets Act, and terrorist-content regulation. That index confirms DSA is one stream among several. It does not, in the fetched listing, reproduce the 112 million figure; the figure comes from the newsroom summary of the sixth DSA report.
This briefing does not use a widely circulated Q1 2026 global Community Guidelines number of 184 million video removals that appeared in some South Asian press. The official Community Guidelines Enforcement URL loaded as a JavaScript application without extractable quarterly tables in the fetched snapshot, and a newsroom URL for a “seventh” DSA report redirected to an unrelated album promotion. Until a loadable official page states Q1 2026 global video totals, those press figures stay out.
Limits of the 112 million: it is an EU DSA perimeter, not a worldwide video-only count. Comments were added in this edition, so the total is not comparable without adjustment to earlier DSA editions that omitted comments. “Around 112 million” is TikTok’s rounding. The 178 million monthly EU users is TikTok’s reach claim for that period. Covert influence and government-request series are separate documents; mixing them with DSA removals would double-count or confuse legal bases.
What this means in practice
Researchers and journalists in Bangladesh can use the sixth DSA report as TikTok’s EU-scale enforcement snapshot for H2 2025, with a clear automation share. They should not treat 93.8% automated as a global accuracy rate, or 112 million as a Bangladesh takedown census. Creators appealing EU removals follow in-app tools TikTok describes on its Transparency Center. Cyber71 cannot fast-track a DSA appeal or obtain unlisted influence-operations indicators.
What is still unclear
The newsroom summary does not break 112 million into videos versus comments versus LIVE. Precision of 97.6% is TikTok’s internal confirmation rate; false-negative rates are not in the loaded post. How TikTok Shop’s 15 June 2026 EU expansions affect later reports is a seventh-report topic that could not be loaded from the official newsroom URL tried for this workstream. Country-level EU member splits live in accompanying CSV files that were not fully parsed here.
Readers comparing TikTok with YouTube or Meta should keep the legal perimeter in mind. DSA numbers count actions TikTok reports for the Union, using TikTok’s policy bundle, over six months, including comments for the first time. YouTube’s quarterly global Community Guidelines tables and Meta’s adversarial-network case studies are different instruments. Adding 112 million to any global video figure from another TikTok product report would be a methodological error.
TikTok’s newsroom framing emphasises investment and automation accuracy. Independent researchers cannot see the labelled sample behind 97.6% “confirmed as correct,” nor the appeal overturn rate for those automated actions. The Transparency Center’s covert-influence and government-request reports, listed on the index page, are the place to look for integrity operations and legal process — not the DSA removal total.
Sources
- TikTok Newsroom — Digital Services Act: Our sixth transparency report on content moderation in Europe2026-09-24
- TikTok — Reports | Transparency Center2026-09-24
Cyber71 links the primary material behind each report so readers can check it directly. External pages are not under our control and may change.
