Platform disclosure guide
TikTok AIGC labels: when and how to disclose
Direct answer
On TikTok, AIGC includes AI-generated or AI-modified image, video and audio. TikTok requires creators to label AI-generated content containing realistic images, audio or video and encourages labeling fully generated or significantly edited content. The creator label and automatic label are different; an automatic label can arise from TikTok effects or Content Credentials. Check the actual published post, especially after scheduling.
What does AIGC mean on TikTok?
AIGC means AI-generated content. TikTok uses the term for generated or AI-modified images, video and audio, including realistic and stylised media. A minor enhancement is different from a significant edit that changes what a person appears to do or say.
When should you use the AI-generated content label?
TikTok requires a label for AI-generated content containing realistic images, audio or video and encourages labels for content fully generated or significantly edited with AI. A creator can use the post setting; TikTok can also auto-label from its effects or Content Credentials. A visible explanation in the video or caption can clarify which component was generated.
Practical decisions
AI voice-over, generated images and edited video: examples
Each case describes a specific final asset. The platform decision is separate from legal duties, rights and client policy.
A realistic AI voice narrates real product footage
Asset: The visuals are camera footage, but the narration is synthetic speech.
Platform decision: TikTok's realistic-audio labeling rule may apply even though the footage is real; the EU deepfake test remains a separate question.
Action: Apply the AI-generated content setting in the native posting workflow and explain that only the narration is generated when useful.
Evidence to keep: Keep the final video, voice-origin record and published label capture.
A fully generated image is posted as a photo
Asset: The image depicts a plausible product scene that never existed.
Platform decision: Realistic generated imagery is within TikTok's required labeling scope; a credential may also prompt an automatic label.
Action: Use the photo post setting for AI-generated content; inspect the final file before upload and verify the post after publication.
Evidence to keep: Keep export hash, metadata observation and screenshot of the photo label.
Inspect embedded AI signals before posting. This does not predict whether a platform will definitely add an AI label.
An existing video makes a person say new words
Asset: AI changes the speaker's speech or voice in otherwise real footage.
Platform decision: TikTok gives altered speech as an example of significant AI editing. Labeling does not resolve impersonation or consent concerns.
Action: Check rights and potential deception first; if publishing is permitted, use the platform label and specific explanatory wording.
Evidence to keep: Retain permission scope, edit history, review decision and the published post.
A third-party scheduler publishes the video
Asset: A campaign system schedules a labeled creative, but its interface does not show a verified AIGC control.
Platform decision: Do not equate a scheduler checkbox or an absent rejection with TikTok's final native label state.
Action: Verify the post after publishing. If the needed native setting cannot be confirmed, use an appropriate direct publishing path and preserve visible wording.
Evidence to keep: Keep scheduler job, post URL, time-stamped label capture and any corrective action.
What should you check after publishing through a scheduler?
Open the published post and verify the actual label or component-specific visible disclosure. The available TikTok documentation describes the native app setting; it does not establish equivalent support for every third-party scheduler. If the necessary state cannot be verified, use a direct publishing workflow and retain the scheduler job and live-post evidence.
Do you need disclosure for your asset?
Check the final media, AI use, audience and destination separately. This is not an inspection of the file's metadata.
First-party policy text was checked on 23 September 2026. These are documented workflows, not a live test of each account, region, post format, ad placement or scheduler interface. Verify the actual publishing UI and served content. See the source register.
Applies when
- Required for AI-generated content containing realistic images, audio or video.
- Trigger scope: Realistic AI-generated or significantly edited media.
- Video: AI-generated content setting under More options in the native video-post flow.
- Image: AI-generated content setting under Post settings in the native photo-post flow.
- Audio: Covered by AIGC labeling when realistic.
Doesn't apply when
- Text: Disclose in caption where relevant.
Worked decision
Example
A realistic AI voice narrates real product footage
Assess the final asset and delivered label
TikTok's realistic-audio labeling rule may apply even though the footage is real; the EU deepfake test remains a separate question.
Next steps
Required action
- For a photo, use Post settings; for a video, use More options in TikTok's documented native flow.
- Turn on the AI-generated content setting where the realistic or significantly edited AIGC rule applies.
- Publish and check whether the creator or automatic label is present on the final post.
- For scheduler posts, verify the live state; if the control is not established, use a direct publishing path.
Publishing layer
Platform differences
EU legal layer
EU AI Act Article 50(4) requires deployers to disclose realistic AI-generated or manipulated image, audio or video resembling real persons, objects, places or events. Article 50 obligations apply from 2 August 2026.
Scheduler and API
Third-party posting interfaces may not expose the AIGC toggle. Verify post-publish or publish natively.
Known limitation
TikTok labels content created with its own AI effects automatically; external tools rely on creator disclosure or metadata detection.
Evidence
Source
Freshness and accountability
Last verified
Last reviewed 2026-09-23 · Current
Run this exact scenario
A representative scenario and TikTok are preselected. Confirm the final media, AI operation and audience.
Native AI-label capability
| Native label | full AI-generated content (AIGC) label |
|---|---|
| Applies to | Required for AI-generated content containing realistic images, audio or video. |
| Surface | TikTok photo and video posts |
| Trigger scope | Realistic AI-generated or significantly edited media |
| Permanence | Platform-controlled |
| Reversible after publication | not verified |
| Rollback path | No verified path recorded |
| Rollback confidence | unknown |
| Post-publish edit | unknown |
| Label state verifiable | not verified |
| Legal threshold | independent Signals TikTok's AIGC classification and must be assessed separately from applicable law. |
| API / scheduler | uncertain Third-party posting interfaces may not expose the AIGC toggle. Verify post-publish or publish natively. |
Content and media types
| Type | Native handling | Manual action |
|---|---|---|
| Video | native | AI-generated content setting under More options in the native video-post flow. |
| Image | native | AI-generated content setting under Post settings in the native photo-post flow. |
| Audio | native | Covered by AIGC labeling when realistic. |
| Text | manual | Disclose in caption where relevant. |
Platform policy can change independently of EU law. Disclara records both layers and never treats a native label as a legal-compliance verdict.