Do You Have to Label AI Videos? Rules by Platform

Yes, if the AI video looks real. YouTube, TikTok and Meta (Facebook, Instagram and Threads) all require a label on realistic AI-generated or AI-altered video, and none requires one for cartoon or clearly fantastical content. YouTube and Meta also cover AI music. Paid promotion needs an ad disclosure on top, under the FTC's Endorsement Guides.

Every quote below comes from the platform's own policy page, read on 2026-10-30. Policies change, so each section links the page to check before you post. This is a practical summary, not legal advice.

What does each platform require?

The table is the short version. The sections after it quote each rule in full.

The trigger everywhere is realism, not the use of AI. Music and voice are where the platforms disagree, so a clip that needs no label on YouTube can still need one on Instagram. Cross-post with the strictest rule.

What does YouTube require?

YouTube requires you to disclose realistic AI content at upload. Its disclosure page says creators "must disclose GenAI content that":

• "Makes a real person appear to say or do something they didn't do."
• "Alters footage of a real event or place."
• "Generates a realistic scene that didn't actually occur."
• "Creates music that's the main focus of the video."

It adds that this "could include content that is fully or partially altered or created using AI tools."

What does not need disclosure on YouTube. The page lists "non-realistic content that's made with AI" and minor edits. Its examples include "Someone riding a unicorn through a fantastical world", beauty and lighting filters, and "Production assistance, like using generative AI tools to create or improve a video outline, script, thumbnail, title, or infographic". It also exempts "Caption creation", "Idea generation" and "Cloning one's own voice to create voice overs or dubs".

What does. YouTube's examples of required disclosure include "AI generated music", "AI generated extra footage of a real place" and "Making it appear as if someone gave advice that they did not actually give".

How to do it. In YouTube Studio, during upload, open the Attributes section and under "AI use" select Yes. YouTube then adds a label. For "AI-generated music or content that is photorealistic", the label may also appear on the video player.

Does it cost you views or money? No, according to YouTube: "Disclosing AI content alone won't limit a video's audience or impact its eligibility to earn money." The risk sits on the other side. Creators who "consistently choose not to disclose" may get a label applied for them, or "removal of content or suspension from the YouTube Partner Program."

Automatic labels. YouTube may label content on its own when it contains C2PA metadata or when "our internal systems detect" AI. You can usually correct a wrong automatic label in Studio. You cannot change one on "content containing C2PA metadata, or content labeled after manual review."

A label does not make a video monetizable. Monetization has its own rules on mass-produced and reused content, covered in what YouTube demonetizes on faceless AI channels.

What does TikTok require?

TikTok requires a label on realistic AI content and lets you choose how. The Community Guidelines say: "we require creators to label AI-generated or significantly edited content that shows realistic-looking scenes or people." The expanded section adds: "You can add your own clear caption, sticker, or watermark. For AI-generated content, you can also use our AIGC label."

TikTok lists cases that need disclosure "when content isn't harmful but could be confusing", including:

• "A face is replaced with someone else's"
• "AI tools make it look like someone said something they didn't"
• "A background, object, or person is added or removed in a misleading way"
• "AI-generated audio mimics the voice of a real person"

And cases that do not:

• "Making small edits like color correction, reframing, or cropping"
• "Using artistic styles, like anime"
• "Using generic text-to-speech (TTS) narration, when the TTS isn't a recognizable voice of a known individual"

If you skip it. "Unlabeled content may be removed, restricted, or labeled by our team, depending on the harm it could cause."

What a label does not fix. Some AI content is banned even with a label. TikTok's not-allowed list includes "Using the likeness of private figures without consent" and AI content that shows "A public figure taking political stances, supporting products, or commenting on public issues they haven't actually addressed". A realistic AI video of a celebrity recommending your product breaks this rule whether or not you label it.

What do Facebook, Instagram and Threads require?

Meta may require a label on realistic AI video and audio, but not on images. Its Help Center says: "On Facebook, Instagram and Threads, Meta may require an AI label when content has photorealistic video or realistic-sounding audio that was digitally created, modified or altered, including with AI. This requirement does not apply to images."

Meta's examples of content that requires a label:

• "A video that appears realistic of a group of people walking around an outdoor market"
• "An audio file of two people talking"
• "A song created using AI-generated vocals"
• "A reel narrated with a realistic AI-generated voiceover"

Examples that do not: a video "created in a style resembling a cartoon", and an "Image of a person riding a bull (Meta does not require a label for images)".

If you skip it. "There may be penalties for content shared without a label when it is required." The page does not say what they are.

Automatic labels. Meta reads "an industry-standard signal" in some AI files, and "images may still receive a label if the systems detect that they were AI-generated". So an AI image can get a label even though you were not required to add one.

Profile label. Accounts that "mostly create and post content that is AI-generated or modified" can add an "AI creator" label to the profile. It is optional on the page we read. For AI personas, our AI influencer disclosure rules by platform covers bios and profile labels in more detail.

The Meta page does not give step-by-step instructions for adding the label in each app, so we have not described the buttons here.

Do you have to disclose AI music?

On YouTube and Meta, usually yes. On TikTok, only in the voice case.

• YouTube. Disclose AI music when it "Creates music that's the main focus of the video". A music video, a lyric video or a song-first Short with an AI track is the main focus. A quiet AI underscore under a talking explainer is less clear-cut. YouTube's example list simply says "AI generated music", so label when in doubt. Music artists get a separate YouTube article.
• Meta. "A song created using AI-generated vocals" needs a label. An instrumental is not named either way.
• TikTok. The guideline names "AI-generated audio mimics the voice of a real person". AI music that does not imitate a real singer is not listed.

Labelling is separate from rights. A label does not clear a song for monetization, and YouTube's reused content policy names "Songs modified to change the pitch or speed, but are otherwise identical to the original song" as not monetizable. Our guide to royalty-free AI music for Reels and TikTok covers the rights side.

Do you have to label an AI voiceover?

The three platforms give three different answers, which makes voice the easiest place to slip.

• YouTube. "Cloning one's own voice to create voice overs or dubs" needs no disclosure. A synthetic voice that makes a real person "appear to say" something does.
• Meta. "A reel narrated with a realistic AI-generated voiceover" needs a label, with no exception for your own voice in the text we read.
• TikTok. Generic TTS needs no label "when the TTS isn't a recognizable voice of a known individual". A voice that "mimics the voice of a real person" does.

So a faceless Reel narrated by a realistic AI voice needs a label on Instagram, even if the same video needs none on YouTube. If you cross-post, the safe setting is to label it everywhere.

Does the FTC require you to label AI content?

The FTC's Endorsement Guides do not require an "AI" label. They require you to disclose paid relationships and keep claims honest, and they apply to AI characters too.

• AI personas count as endorsers. The guides define the endorser as a party that "could be or appear to be an individual, group, or institution." When the FTC adopted that wording in 2023, it said the language is meant "to also encompass the writers of fake reviews and non-existent entities that purport to give endorsements."
• Paid relationships must be disclosed. A connection that is "not reasonably expected by the audience" must be "disclosed clearly and conspicuously". That includes payment and free product.
• Fake customers must be disclosed. Ads that present "actual consumers" should use real consumers "or clearly and conspicuously disclose that the persons in such advertisements are not actual consumers of the advertised product." An AI-made "customer testimonial" falls here.
• The platform's tool may not be enough. One FTC example describes a creator who "relies solely upon a social media platform's built-in disclosure tool". The disclosure was in small white text on a light background, competing with other text, and the post appeared for only five seconds, so it was "easy to miss and thus not clear and conspicuous."
• Say it where the claim is made. If the endorsement is spoken and shown, the disclosure should be "in the communication's visual and audible portions."

For ad work for brands, our AI UGC disclosure checklist has example wording.

How do you label an AI video before you post?

Run these steps on every post. They take a few minutes once the habit is set.

1. List every AI part of the video. Write down which shots, voices and music are AI-made and which are real footage. If you made them on Oakgen, each file in your Assets library shows its prompt, what it was made with and the date, so the list is a quick lookup. The parts usually come from the AI video generator, voice and text-to-speech and the AI music generator.
2. Mark each part as realistic or not. Realistic people, places and events go in the "label" column. Cartoon, anime and obvious fantasy go in the "no label" column on all three platforms.
3. Check the voice and music lines separately. Use the voiceover and music sections above. A realistic AI voice is labelled on Meta even when YouTube does not require it.
4. Set the label on each platform. On YouTube, set "AI use" to Yes under Attributes at upload. On TikTok, turn on the AIGC label or add a clear caption, sticker or watermark. On Facebook and Instagram, add the AI label when you share.
5. Add the ad disclosure if anyone paid you or sent you product. Put it on screen and say it out loud if the endorsement is spoken. Do not rely on the platform's paid-partnership tag alone.
6. Check what a label cannot fix. No real person's face or voice without consent, no fake news footage, no public figure "endorsing" your product. These break platform rules whether or not they are labelled.
7. Keep the record. Save your list from step 1 with the post link. If a platform questions a label later, you can show what was AI and what was not.

What can Oakgen do here, and what can it not?

Oakgen makes the AI parts of a video. It does not label them for you, and its video tools do not upload to YouTube, TikTok or Instagram. We checked the code on 2026-10-30. Oakgen does not write its own C2PA content credentials into files, and we have not checked which model providers embed their own metadata. Assume any platform may detect AI in your file and set the label yourself.

Whether you can use an output commercially depends on section 4.4 of Oakgen's Terms. Oakgen does not claim ownership of your inputs or eligible outputs. Your right to use an output depends on the Terms, the law, the model provider's terms and the rights you hold in any faces, voices, logos or other material you use. A platform label does not change any of that.

How we checked this

We read each platform's policy page on 2026-10-30 and copied the quotes from the page text: YouTube's GenAI disclosure page and channel monetization policies, the Integrity and Authenticity section of TikTok's Community Guidelines (including its expanded "More information" lists), and Meta's Help Center page on identifying AI content. For the FTC, we read 16 CFR Part 255 on eCFR, shown as up to date as of 2026-10-28, and the 2023 Federal Register notice that revised the endorser definition. Links are in the sources list.

We did not test how each app's label button looks or behaves, and we did not check country-specific laws such as the EU AI Act. The Oakgen product details come from Oakgen's code on the same date. Platform rules change often, so check the linked page before a high-stakes post.

Frequently asked questions

Do I have to label every AI video?

No. YouTube, TikTok and Meta all tie the rule to realism. A cartoon, anime or clearly fantastical AI video does not need the label on any of the three, based on their policy pages read on 2026-10-30. A realistic person, place or event that did not happen does need it.

Do I have to disclose AI music on YouTube?

Yes, when the AI music is the main focus of the video. YouTube's disclosure page lists content that "Creates music that's the main focus of the video" and gives "AI generated music" as an example. Meta also lists "A song created using AI-generated vocals" as content that needs a label.

Does an AI voiceover need a label?

It depends on the platform. YouTube says cloning your own voice for voiceovers does not need disclosure. Meta lists "A reel narrated with a realistic AI-generated voiceover" as needing a label. TikTok says generic text-to-speech that is not a recognizable voice of a known person does not need one. If you post to all three, label it.

Does the AI label hurt views or monetization?

YouTube says "Disclosing AI content alone won't limit a video's audience or impact its eligibility to earn money." TikTok and Meta pages do not make that promise either way. Separate originality rules, such as YouTube's monetization policies and TikTok's For You feed rules, can still cut reach or pay for mass-produced or reused content.

Does the FTC require an AI label?

The FTC's Endorsement Guides are about paid relationships and honest claims, not AI as such. But an AI persona that promotes a product is still an endorser under the definition, which covers a party that could "be or appear to be" an individual. Paid promotion needs a clear disclosure whether a person or an AI character delivers it.

What happens if a platform labels my video automatically?

YouTube may add a label when it detects C2PA metadata or AI content. You can usually change a wrong label in YouTube Studio, but not one based on C2PA metadata or a manual review. Meta reads industry-standard signals in files and may label images too, even though it does not require you to label images.

Sources

1. YouTube Help, Disclosing use of GenAI content (read 2026-10-30)
2. YouTube Help, YouTube channel monetization policies (read 2026-10-30)
3. TikTok Community Guidelines, Integrity and Authenticity (read 2026-10-30)
4. Meta Help Center, How to identify AI content on Meta products (read 2026-10-30)
5. eCFR, 16 CFR Part 255, Guides Concerning the Use of Endorsements and Testimonials in Advertising (up to date as of 2026-10-28, read 2026-10-30)
6. Federal Register, 88 FR 48092, FTC Endorsement Guides final revision, 26 July 2023 (read 2026-10-30)