TikTok removed 14,791,485 videos in Bangladesh between April and June 2026 for breaking its community guidelines, according to its latest enforcement report, released on 29 October. The figure is about 0.9 per cent of the content uploaded from the country in that period.
Almost all of the removals happened without a user complaint. TikTok says 99.8 per cent of the videos taken down in Bangladesh were caught proactively, meaning before anyone reported them. Another 205,001 videos were later restored after review.
How Bangladesh compares
Globally, the company reported removing more than 188.3 million videos in the quarter, around 0.5 per cent of everything uploaded. Of those, 183.3 million were caught by automated detection, and the proactive removal rate was 99.5 per cent. TikTok says roughly 92.7 per cent of proactively removed videos came down within 24 hours of being posted.
Bangladesh’s share of removals is higher than its share of the world’s uploads would suggest. The 0.9 per cent removal rate here is nearly double the global rate, and the country accounts for roughly 8 per cent of all videos removed worldwide. The report does not explain the gap.
The global breakdown shows where most violations occur. Sensitive or mature themes made up 19.5 per cent of removed videos, body exposure and sexualised behaviour 13.8 per cent, safety and civility violations 13.7 per cent, and dangerous activities and challenges 13.5 per cent. Edited media and AI-generated content accounted for 1.8 per cent.
Counting removals is not the same as measuring harm. A high figure can point to effective enforcement, to heavy rule-breaking, or to both, and the report alone cannot tell which. It is best read as a gauge of scale and of how much work automated systems now do.
TikTok also removed more than 278.3 million comments globally, up from 237.5 million in the previous quarter. More than 6.2 million videos were reinstated after review worldwide.
What it means here
These are figures published by the platform itself, and they cannot be checked independently. The report does not say how well automated systems handle Bangla, regional dialects or local context, which is where moderation errors are most likely to appear.
The restored videos offer one rough measure. In Bangladesh, about 1.4 per cent of removed videos were later reinstated, which suggests that most decisions stood but that some creators were wrongly caught. Creators who depend on the platform for income have a clear interest in an easy appeals process.
Users have a part to play as well. Creators whose videos are removed can appeal, and the reinstated count shows that appeals sometimes succeed, so it is worth keeping records of what was posted and when.
As Bangladesh debates how to regulate social media and AI-generated content, enforcement reports like this one give policymakers a baseline. The next step is pressing platforms for more detail on local-language performance and appeals, so that the numbers say as much about fairness as they do about scale.
