Can viewers see less AI video on TikTok? The AIGC control

Yes, in testing. TikTok's newsroom post dated March 10, 2026 says it is testing a Manage Topics AIGC control that lets people choose how much AI-generated content they see in For You (read 2026-10-03). TikTok says it is intended to tailor the feed, not remove content entirely.

For creators, the practical meaning is that an accurate AI label may decide which viewers see a clip. That is a reason to label correctly, not to avoid labels.

What does TikTok say about the control?

The post says Manage Topics already covers over 10 categories, and that the AIGC control is being tested within it. It also says TikTok is testing invisible watermarking, because labels can be removed when a video is re-uploaded or edited, and that watermarks are being added to content made with TikTok AI Editor Pro and to content uploaded with C2PA Content Credentials, readable only by TikTok. TikTok adds context on whether content was labeled due to AI detection, creator labels, or TikTok AI tools (read 2026-10-03).

What does the control mean for AI video reach?

TikTok does not publish reach numbers for the control, and we did not find any in the pages we read. Anything you see quoted about reach loss is a third party's observation. Our post on what TikTok and Instagram reports say about AI labels and reach keeps those sources labeled as reports.

The control being a viewer choice is the useful fact. It means a labeled clip can still be shown to people who want AI content, and an unlabeled one that TikTok later detects is shown with a label attached. Hiding the label buys nothing a viewer cannot undo.

What should an AI video creator do now?

Treat the label as audience targeting, and keep the production record that supports it. Sume produces and edits the files; the label itself is set in TikTok at upload. Sume's job record holds the job type, model when available, timestamps and artifact URL, which is the part you can prove.

• Label every clip that shows realistic-appearing people or scenes, per TikTok's rule.
• Group your AI series under consistent topics so viewers who opt in find them.
• Record the label state at upload in your tracker next to the Sume job id.
• Watch TikTok's own pages for the control's rollout; it is described as a test.

Does a label prevent the For You feed from showing the clip?

TikTok's post says the control tailors the feed rather than removing content entirely. So a viewer who turns AIGC down will see less of it, not none. It is an input to ranking for that viewer, and the page gives no mechanism beyond that.

Do not confuse this with the separate rule on reused content. Batches of near-identical clips raise a different problem, covered in TikTok's unoriginal content rule for AI video batches. The AIGC control is about the kind of content, not how many copies exist.

What should you not conclude from a test?

TikTok describes the control as a test. Tests change, narrow or disappear, and the page does not say when or where it reaches everyone. Do not plan a content strategy on the assumption that the control exists for all your viewers.

Also avoid the opposite mistake of treating it as a threat. The page frames it as tailoring, and a viewer who opts into AI content may actively look for it. Accurate labels help those viewers find you. What hurts is a mismatch: an unlabeled generated person, or a label on a clip that is entirely filmed.

The steady move is the boring one: label what is generated, keep the record, and recheck TikTok's own pages monthly.

How do you keep up with changes?

Bookmark TikTok's newsroom and transparency page and note the read date in your tracker each time you check them. The March 2026 post says several items are in test, so the wording will likely move. When it does, update your internal label policy once and apply it to every clip in the pipeline rather than patching individual posts.