TikTok’s AI disclosure rules apply to content made or significantly edited with AI when it depicts realistic-looking people or scenes. The requirement covers images, video and audio, including voice clones and clips altered to make someone appear to say something they did not say.
The platform distinguishes these changes from minor edits. Color correction, cropping, reframing and everyday filters such as makeup effects do not need an AI label. Generic text-to-speech voiceovers and AI-written captions or scripts are also listed as not requiring disclosure.
For organic posts, creators can turn on the AI-generated content setting in the app before publishing. TikTok may also add a label automatically, including when a post uses its AI effects or carries Content Credentials metadata. Creators can additionally disclose AI use in a caption, sticker or watermark.
Advertisers must use a separate setting in TikTok Ads Manager for ads containing AI-generated or significantly AI-edited realistic elements. The source guide says Spark Ads follow the disclosure process for organic posts, so the original post should be labeled correctly.
If a post is not labeled, TikTok may add a label, limit its reach or remove it, depending on the potential for harm. Ads with undisclosed AI content may also be rejected or restricted. The guide notes that adding a label does not, by itself, reduce a post’s reach.
TikTok also prohibits some AI content regardless of whether it is labeled, including material that could seriously mislead people about public issues or harm a real person. For branded content, the guide recommends checking the platform’s latest policies and consulting legal teams; an AI label does not replace disclosure of a paid creator partnership.
Sources: Hootsuite Blog.