The TikTok AI Generated Label Is Two Different Labels

What is the TikTok AI generated label?

Two separate mechanisms produce a TikTok AI generated label: one the creator switches on, and one TikTok applies by itself. That is why creators keep asking why a label showed up on a post they never marked. TikTok's help article "About AI-generated content" describes both under one heading, and the two labels do not say the same thing.

The first is the creator label. TikTok's wording: "Creators apply this label to indicate that their content was completely generated or significantly edited by AI. The label will read "creator labeled as AI-generated"." You turn it on, and you can turn it off.

The second is the auto label. TikTok's wording again: "TikTok may automatically apply the "AI-generated" label to content we identify as completely generated or significantly edited with AI." Nobody asks you first.

So the visible text differs. One reads creator labeled as AI-generated, the other reads AI-generated. Compliance checklists written from secondhand blog posts collapse the two into a single "AI label," and then the person maintaining the checklist cannot explain why two videos on the same account show different words.

Where do you turn on the AI-generated content setting?

Inside the TikTok app, on the post screen, and the toggle is not named after the label it produces. TikTok calls the control the AI-generated content setting. The label it produces reads creator labeled as AI-generated. Searching the app's settings for the label text finds nothing, because the label text exists only on the published post.

The documented paths differ by media type. For a photo, tap the Add post + button, capture the photo, tap the Post settings at the bottom, and turn the AI-generated content setting on. For a video, tap Add post +, record and tap Continue, or upload from your device, tap Next, then tap More options and turn the setting on there. The video path buries it one level deeper than the photo path, which accounts for a good share of the confusion.

When does TikTok apply the AI-generated label automatically?

TikTok documents two triggers plus its own detection. The help article names them in one sentence: "This may happen when a creator uses TikTok AI effects or uploads AI-generated content that has Content Credentials attached, a technology from the Coalition for Content Provenance and Authenticity (C2PA)." It adds that "Content Credentials attach metadata to content that we can use to recognize and label AIGC instantly."

The C2PA path is the one that catches people. Content Credentials travel inside the file's metadata, written by the tool that generated the image or video. Many generators write them by default. Upload that file and the label attaches on the way in. You do not opt in, you are not prompted, and nothing in your posting flow signals it will happen.

The TikTok effects path works in your favour, and TikTok says so directly: "if you made your AI-generated content using only TikTok effects, you may not need to label it as we automatically label content made with TikTok effects if they use AI." The exception follows immediately. "If you independently altered your video with AI in addition to the TikTok effect you used, we still ask that you follow our labeling guidelines." An AI effect plus your own AI edit puts the manual label back on you.

Can you remove the AI-generated label once TikTok applies it?

No, and TikTok states it as a flat rule with no conditions: "Note: Once your content is labeled as AI-generated with an auto label, you won't be able to remove the label from your post." There is no toggle, no cooldown, no second chance on the same post.

TikTok does not publish an appeal route for an auto label the creator believes is wrong. The reporting flow documented in the same article is for reporting somebody else's video under the misinformation policy, not for disputing a label on your own. If a route exists, TikTok does not document it in this article.

YouTube runs the same irreversibility rule with different words, calling the creator control the "AI use" setting under Attributes in YouTube Studio and stating that content made with YouTube's AI tools, content containing C2PA metadata, or content labeled after manual review "can not be adjusted," so a label workflow copied from TikTok gets both the setting name and the label string wrong on YouTube.

The practical consequence for anyone scheduling AI assisted content is that the decision happens before upload, not after. Once a file carrying Content Credentials reaches TikTok, deleting and reposting is the only lever, and it costs you the post's entire engagement history. Strip or keep the credentials deliberately, before the file enters the queue.

Which field declares AI-generated content in the TikTok Content Posting API?

The field is is_aigc, a bool, and it is optional on every endpoint that has it. Where it goes in the request body depends on which endpoint you are calling, and TikTok's own reference is inconsistent about both placement and the label it promises.

For a direct video post, is_aigc is a sibling of title and privacy_level:

{
  "post_info": {
    "title": "how I made this #ai",
    "privacy_level": "PUBLIC_TO_EVERYONE",
    "brand_content_toggle": false,
    "brand_organic_toggle": false,
    "is_aigc": true
  },
  "source_info": {
    "source": "PULL_FROM_URL",
    "video_url": "https://example.verified.domain.com/example_video.mp4"
  }
}

For a photo post it climbs out of post_info and sits next to media_type:

{
  "media_type": "PHOTO",
  "post_mode": "DIRECT_POST",
  "is_aigc": true,
  "post_info": {
    "title": "funny cat",
    "privacy_level": "PUBLIC_TO_EVERYONE",
    "brand_content_toggle": false,
    "brand_organic_toggle": false
  },
  "source_info": {
    "source": "PULL_FROM_URL",
    "photo_cover_index": 0,
    "photo_images": ["https://example.verified.domain.com/01.webp"]
  }
}

Why does is_aigc move between the video and photo endpoints?

TikTok gives no reason. Both reference pages carry a "Last updated August 24, 2026" stamp, both were written by the same team, and they disagree with each other in two places at once.

The placement is the first disagreement. Every other post option on the photo endpoint carries the qualifier "Only works for post_mode = DIRECT POST" and lives inside the Post Info Object. is_aigc carries no such qualifier and lives outside it. The reference does not say what happens to an is_aigc sent in the wrong object, and neither endpoint's error table lists a code for it. The documented 400 codes are invalid_param for both, plus app_version_check_failed on the photo endpoint.

The promised label is the second disagreement, and it matters more. The video reference promises the creator string: "labelled with Creator labeled as AI-generated tag in video's description." The photo reference promises "an AI-generated tag in the description," which is the auto label string, not the creator one. Read literally, the same boolean produces two different labels depending on media type. TikTok does not reconcile the two sentences anywhere in the reference, and the help article does not mention the API at all. Test the string on a private post before you promise a client anything about photo posts.

What about the upload path, where the creator finishes in the app?

The declaration is not yours to make. /v2/post/publish/inbox/video/init/ accepts a source_info object and nothing else, so there is no field through which your client can pre-set the label. Your app sends the file, the creator gets an inbox notification, and the creator turns the AI-generated content setting on inside TikTok, or does not.

That path carries its own ceilings worth knowing before you design around it. TikTok caps each user access token at 6 requests per minute across these endpoints, and returns spam_risk_too_many_pending_share once an account has 5 pending shares inside any 24 hour period. The photo endpoint's MEDIA_UPLOAD mode also requires the creator's TikTok app to be version 31.8 or newer, and returns app_version_check_failed when it is not.

Does labelling a post as AI generated cost you reach?

TikTok says no, with one condition attached: "Turning on the AI-generated content setting won't affect the distribution of your video as long as it doesn't violate our Community Guidelines." That statement covers the creator label only. TikTok publishes nothing about how the auto label affects distribution.

Not labelling carries the heavier risk. TikTok requires creators to label all AI-generated content containing realistic images, audio and video under its Community Guidelines, and says it "may remove content that violates our Community Guidelines policies by depicting misleading information or unlabeled AI-generated content." Both directions end in removal. Mislabelling real footage as AI breaks the Terms of Service, and leaving realistic AI footage unlabelled breaks the Community Guidelines.

The label rules belong to TikTok, not to whatever tool sends the post. AdaptlyPost schedules TikTok posts through the TikTok posting API, publishes the same item to several networks in one action, and reads results back through TikTok analytics. Where a scheduler helps is upstream of the label decision: an AI generated image sitting in a content calendar is a file you can still inspect and decide about, and that is the last moment the decision is still reversible.

Frequently Asked Questions

What does the TikTok AI label actually say?

One of two strings. The creator applied label reads "creator labeled as AI-generated". The automatic label reads "AI-generated". TikTok's help article gives both strings verbatim and treats them as separate labels with separate triggers.

Does is_aigc go inside post_info?

For video, yes. post_info.is_aigc is the documented location for /v2/post/publish/video/init/. For photos it does not: the reference for /v2/post/publish/content/init/ puts is_aigc at the top level of the request body, as a sibling of post_info, source_info, media_type and post_mode.

Can I remove an AI-generated label that TikTok applied automatically?

No. TikTok's wording is "Once your content is labeled as AI-generated with an auto label, you won't be able to remove the label from your post." TikTok does not document an appeal path for an auto label applied in error.

Do I need to label content made with TikTok's own AI effects?

Not if the effect was the only AI involved. TikTok automatically labels content made with its AI effects, so it says "you may not need to label it" yourself. If you altered the video with AI separately from the TikTok effect, TikTok still asks you to apply the label.

Will the AI-generated label reduce my views?

TikTok says the creator label does not, stating that turning the setting on "won't affect the distribution of your video as long as it doesn't violate our Community Guidelines." TikTok publishes no equivalent statement about the automatic label.

Can an API client set the AI label on the inbox upload endpoint?

No. The body schema for /v2/post/publish/inbox/video/init/ documents source_info only, with no is_aigc field. On that path the creator applies the label in the TikTok app after tapping the inbox notification.

Why does TikTok show an AI-generated label you never added?

TikTok attaches the auto label after reading either its own AI effects or Content Credentials, the C2PA metadata that many AI generators write into a file by default. Upload that file and the label attaches on the way in, with no opt-in prompt and no setting to check first. Nothing in your posting flow signals that it will happen, which is why the label appears on a post you never marked.

Does YouTube handle AI labeling the same way TikTok does?

Not exactly. YouTube calls its creator control the AI use setting under Attributes in YouTube Studio, a different name for a similar toggle. Content made with YouTube's AI tools, content carrying C2PA metadata, or content labeled after manual review all end up unable to be adjusted, the same irreversibility rule TikTok applies to its auto label, just worded differently. A label workflow copied straight from TikTok gets both the setting name and the label string wrong on YouTube.

What happens if you skip labeling AI-generated content on TikTok?

TikTok's Community Guidelines require creators to label AI-generated content containing realistic images, audio, or video, and the platform may remove content for depicting misleading information or unlabeled AIGC. The risk runs both directions: mislabelling real footage as AI breaks the Terms of Service, while leaving realistic AI footage unlabelled breaks the Community Guidelines. Either mistake can end with the post removed.

What are the rate limits on the TikTok Content Posting API?

TikTok caps each user access token at 6 requests per minute across the posting endpoints. An account that racks up 5 pending shares inside any 24 hour period gets spam_risk_too_many_pending_share back instead of a successful post. The photo endpoint's MEDIA_UPLOAD mode adds its own gate too, requiring the creator's TikTok app to be version 31.8 or newer and returning app_version_check_failed when it isn't.