// HOW-TO · AI DISCLOSURE

How to disclose AI-generated content on TikTok (2026)

How to disclose AI-generated content on TikTok: when the label is required, the organic AI toggle steps, the Ads Manager switch, Spark Ads, and auto-labeling.

Last verified · 2026-11-03 · by Moe Ameen

TikTok's AI disclosure rule is simpler to comply with than it is to understand. If you post AI-generated or AI-edited content that looks realistic — a synthetic person, a cloned voice, a convincing scene that didn't happen — you have to flag it, and TikTok gives you one switch to do it. The catch is knowing which of your content actually trips the rule and using the right control for organic posts versus ads, because the two paths are separate and the ad path has traps that quietly un-label your creative.

This is the hands-on version, in order: decide whether a given piece even needs the label, turn on the organic "AI-generated content" toggle in the composer, use an in-content disclosure as a backup when you want it visible in the frame, flip the separate switch for a paid ad in Ads Manager, handle the Spark Ads and branded-content cases, and understand what TikTok labels automatically so you don't second-guess a clean upload. For the full policy — the exact realism line, the exemptions, and the enforcement behind each step — read the guide [TikTok AI disclosure policy](/guides/tiktok-ai-disclosure-policy). If your concern is the government rules that sit on top of the platform's, see [how to comply with state AI-ad disclosure laws](/how-to/comply-with-ai-ad-disclosure-laws).

The steps

Decide whether this piece actually needs the label. Run the realism test before you reach for the toggle. Label it if the content is AI-made or AI-edited and could pass as real: a synthetic or face-swapped person, a cloned voice of a real individual, changing what someone said, or adding/removing people or objects in a misleading way. Do not label purely cosmetic or textual AI — color correction, cropping, everyday beauty filters, generic text-to-speech, AI-written captions or scripts. The question is "could a viewer think this is real?", not "did AI touch the file?"
Turn on the organic AI-generated content toggle. For a normal post, the disclosure is four taps. Create or upload your video or slideshow and tap Next to reach the post screen. Tap More options. Turn on the control labeled "AI-generated content." Publish. TikTok then attaches a visible label near your username reading along the lines of "Creator labeled as AI-generated," so viewers see the disclosure the same way they see your caption. There is no penalty for labeling a borderline piece, so when a clip sits on the line, flip it on.
Optionally add an in-content disclosure as a visible backup. TikTok also accepts a clear disclosure inside the content itself — a caption line, a sticker, or an on-screen watermark stating the content is AI-generated. This is not a replacement for the toggle on realistic content, but it is useful when you want the disclosure unmistakable in the frame (for example on a synthetic spokesperson) or when you're reposting elsewhere and want the statement to travel with the file. Use it alongside the toggle, not instead of it.
Disclose AI in a paid ad through Ads Manager. Ads use a different control. In TikTok Ads Manager, create or open your campaign, go to the ad creation section, and turn on the switch worded "This ad contains AI-generated content," then finish setup and submit. TikTok's advertising policies require it for a qualifying ad — one whose image, video, or audio is fully AI-generated or significantly modified by AI — before it can run. Important: once you submit, the toggle locks, and the only way to turn it off afterward is to replace the video or image.
Handle Spark Ads and branded content correctly. Spark Ads boost an existing organic post rather than uploading new creative, so they follow the organic disclosure process: the AI label has to be on the underlying organic post before you authorize it for promotion — a label missed there follows the ad into paid distribution. For sponsored/branded content, enable the AI-generated label alongside the commercial-content disclosure so both obligations are covered on the same post.
Know what TikTok labels automatically — and that you can't undo it. Don't assume an unlabeled upload is clean. Content made with TikTok's own AI effects is labeled automatically, and since May 2024 TikTok reads C2PA Content Credentials and auto-labels files made in other AI tools the moment they're uploaded — the credentials persist even through download and re-upload. Once TikTok applies an automatic label, you cannot remove it. So self-disclosure is always an addition to the automatic layer, never a workaround for it.
Verify the label landed before you move on. After publishing, check the live post: the AI-generated label should appear near your handle and caption. For an ad, confirm the toggle reads on in Ads Manager before the campaign goes to review. If you duplicated a campaign, re-check the toggle specifically — duplication resets it to off, which is the single most common way a previously-compliant ad ships unlabeled.

Common gotchas

Labeling the wrong thing. Cosmetic edits, filters, generic text-to-speech, and AI-written captions do not need the label — the rule targets realistic, potentially-misleading AI, not every file AI touched.
Assuming the ad toggle behaves like the organic one. It doesn't: it locks on submit (only swappable media unlocks it) and resets to off when you duplicate a campaign.
Boosting an unlabeled organic post as a Spark Ad. The label must already be on the original post; adding it later won't retroactively cover the promotion.
Thinking you can remove an automatic label. Once TikTok auto-labels a post via Content Credentials or its own AI tools, the creator can't take it off.
Treating disclosure as a universal permission slip. AI that seriously misleads on public issues or harms a real person is prohibited even with the label on.
Confusing platform policy with the law. State and EU AI-ad disclosure rules are separate obligations that can apply on top of TikTok's toggle.

Legal note

TikTok's AI label is a platform policy, enforced through reduced reach, removal, or ad rejection — not a government fine. It is distinct from statutory AI-disclosure requirements (various U.S. state ad-disclosure laws and the EU's transparency rules), which can impose their own obligations on AI-generated advertising regardless of whether you've toggled TikTok's label. Flipping the TikTok switch satisfies TikTok; it does not automatically satisfy the law. Confirm your current obligations against TikTok's live Community Guidelines and Advertising Policies and any applicable regulation before relying on this for compliance.

Where Kompozy fits

Look back at step one — the realism test — and notice it is the only judgment call in the whole process. Steps two through seven are mechanical: a four-tap toggle, an Ads Manager switch, a verification. The actual work of TikTok compliance was never the label. It was producing the realistic AI content in the first place, and then remembering to label it among dozens of other posts. [Kompozy](/) is a full AI content generation and multi-platform publishing engine — not a repurposing tool — and it is the thing doing that production: [Persona Shorts](/glossary/persona-shorts) and Persona HeyGen put a synthetic spokesperson on camera, [Persona Photos](/glossary/output-buckets) render a face-locked AI character — exactly the realistic output that answers step one with "yes, label it."

What changes when Kompozy makes the content is that the labeling decision stops being something you reconstruct after the fact. Because every Kompozy persona runs off one consistent [Persona Brief](/glossary/persona-brief), your AI character is a known, recurring entity — you decide once that this persona's video is realistic AI, and that answer holds for every clip it ever appears in, rather than being re-litigated per post. And because Kompozy generates the whole slate fast — avatar video, carousels, blogs, newsletters from one source — the only step it hands back to you is the 10-second toggle this tutorial covers. Production, which was the real bottleneck, is automated; disclosure, which is trivial once you know the answer, stays in your hands where the policy wants it. You still flip TikTok's switch yourself; Kompozy just makes sure you always know which pieces need it. Paid plans start at the Starter tier and scale through Pro to Enterprise as your output volume grows.

Frequently asked questions

Where is the AI label toggle on TikTok?

For organic posts it's in the post composer: after you tap Next, open More options and turn on "AI-generated content," then publish. For ads it's a separate switch, "This ad contains AI-generated content," inside the ad creation section of TikTok Ads Manager. The two are different controls on different surfaces — the organic toggle does not cover a paid ad, and vice versa.

Do I have to label AI-written captions or an AI voiceover?

Generally no. TikTok's exemptions cover AI-written captions and scripts and generic text-to-speech voiceovers, because they don't create a realistic depiction of something that didn't happen. The label is for realistic synthetic or heavily altered media — a fake-looking real person, a cloned voice of a specific real individual, a face swap, a misleading edit. If your AI use is cosmetic or textual, you don't need the toggle.

Will labeling my content as AI reduce its reach?

Applying the correct label is how you avoid a reach penalty, not cause one. TikTok reduces reach on realistic AI content it detects as undisclosed, and can label or remove it. The bigger risk is leaving the label off content that needed it. Some creators do over-label cosmetic edits unnecessarily, which is why the realism test matters — label the realistic AI, skip the filters and captions, and you signal accurately.

Does disclosing on TikTok cover me for AI-ad disclosure laws?

No. TikTok's toggle satisfies TikTok's policy. Government rules — several U.S. state advertising-disclosure laws and EU transparency requirements — are separate obligations that can apply to AI-generated ads on their own terms. Treat the platform label and the legal disclosure as two different boxes to check; see [how to comply with state AI-ad disclosure laws](/how-to/comply-with-ai-ad-disclosure-laws).