What counts as significantly edited by AI on TikTok ads?

TikTok's advertising policy defines "significantly modified by AI" as content changed by AI beyond minor tweaks or enhancements. Fully AI-generated images, video or audio count. Adjusting lighting, brightness or color saturation, removing or modifying backgrounds, and denoising are listed as insignificant.

The definitions below are from TikTok's Misleading and false content advertising policy, marked "Last updated: April 2026" and read 2026-09-29. The mapping to Sume operations is mine: the page names none of them.

What does TikTok list as significant and insignificant?

Significantly edited media and AIGC is allowed if you apply the AIGC label, or add a clear disclaimer, caption, watermark or sticker of your own. The page lists these examples:

Which Sume outputs land on which side?

Sume's endpoints produce different kinds of files, so the answer depends on the operation. This is my reading of TikTok's wording, not a TikTok ruling.

What if I am not sure?

TikTok's page says that if you are uncertain whether content falls under the significant category, it is best to use caution and label it accordingly. It also says that if TikTok identifies AI-generated content that has not been disclosed, the ad will be rejected or restricted.

A mixed ad is where the question bites. A real product clip with only a Sume background removal is the insignificant case on TikTok's list. TikTok's list puts completely AI-generated video or audio on the significant side, so an ad that adds a generated scene or a synthetic voice falls under it in my reading.

Where do I set the label?

TikTok's page says to apply the AIGC label, or to add a clear disclaimer, caption, watermark or sticker of your own. It does not give a menu path, so check TikTok's current ad-creation screen for where the label lives. Neither of the Sume pages cited here (Video generation and Video filter) describes a label option, so the disclosure step stays with you.