September 24, 2026 · Reachira team · 7 min read

Using AI UGC creators legally: consent, disclosure, labels

Likeness consent, ad disclosure and the AI label rules on TikTok, Instagram and YouTube for brands posting AI UGC creators, with links to each policy.

An AI UGC creator is a realistic presenter, generated or trained from images, who talks about your product the way a real customer might. It sits where three sets of rules meet: rights to a real person's face and voice, advertising law on endorsements, and each platform's own AI-labeling policy.

This article covers all three, with a source for each rule. It is not legal advice. Laws differ by country and change often. If you are running paid campaigns or using a real person's likeness, talk to a lawyer where you operate.

If a character is built from photos of a real person, or speaks with a clone of a real person's voice, you need that person's permission.

The platforms treat likeness broadly. TikTok's Integrity and Authenticity guidelines define likeness as a recognizable image, video or audio representation of a person, including face, body, voice and gestures. They list "using the likeness of private figures without consent" as not allowed.

YouTube lets people act on it. Under its privacy guidelines, someone can ask YouTube to remove content that shows a realistic altered or synthetic version of their likeness. YouTube considers whether the content is disclosed as synthetic, whether the person can be uniquely identified, whether it is realistic, and whether it shows a well-known person doing something sensitive, such as endorsing a product.

What a sensible consent record covers:

Scope: face, voice or both, and whether the voice may be cloned.
Use: organic posts, paid ads, which platforms and which products.
Duration and withdrawal: how long the consent lasts and what happens to published videos if the person withdraws it.
Payment, if any, which also matters for the advertising disclosure below.

Fully synthetic characters avoid the likeness question, but not the two below.

2. Advertising law: the presenter is not a customer

In the US, the FTC's Endorsement Guides govern endorsements and testimonials in advertising. Three provisions matter most for AI UGC:

Endorsements must reflect the honest opinions, findings, beliefs or experience of the endorser (§ 255.1(a)).
When an ad says the endorser uses the product, the endorser must have been a bona fide user of it (§ 255.1(c)).
Ads that present endorsers as "actual consumers" should either use actual consumers or clearly and conspicuously disclose that they are not (§ 255.2(c)).

The FTC's consumer reviews and testimonials rule goes further. It makes it an unfair or deceptive practice for a business to create a testimonial that materially misrepresents that the testimonialist exists, or that they used the product (§ 465.2).

So an AI character can present and demonstrate your product, but should not claim experiences it could not have had, like "I lost ten pounds" or "this saved my business", unless the video makes clear it is not a real customer.

The Guides define "clear and conspicuous" as difficult to miss and easy to understand. If the claim is visual, the disclosure should be at least visual, and if the claim is spoken, the disclosure should be at least spoken. The FTC's Endorsement Guides FAQ adds that a platform's disclosure tool is not automatically enough: responsibility rests with the influencer and the brand.

TikTok's commercial disclosure policy requires the content disclosure setting when you promote your own business or post incentivized branded content. It says undisclosed commercial content may have its visibility reduced, and that repeated failures can lead to posting restrictions or an account ban.

In the EU, the AI Act (Article 50) requires deployers of an AI system that generates or manipulates a "deep fake" to disclose that the content is artificially generated or manipulated. The Act defines a deep fake as AI content that resembles existing persons, objects, places or events and would falsely appear authentic. The Act has since been amended, so check the consolidated text for when this applies.

3. The AI label on each platform

TikTok

TikTok's Integrity and Authenticity guidelines require creators to label AI-generated or significantly edited content that shows realistic-looking people or scenes. You can use TikTok's AIGC label or your own clear caption, sticker or watermark. Disclosure is needed when, among other cases:

A face is replaced with someone else's.
AI makes it look like someone said something they didn't.
AI-generated audio mimics a real person's voice.

It is not needed for small edits like colour correction or cropping, for clearly artistic styles like anime, or for generic text-to-speech that is not a recognizable person's voice.

TikTok also labels content automatically. In May 2024 it began reading C2PA Content Credentials to auto-label AI content uploaded from other platforms. Publishing tools can use the Direct Post API is_aigc field, which shows a "Creator labeled as AI-generated" tag.

Instagram

Meta's AI labeling help page requires a label on photorealistic video or realistic-sounding audio that was digitally generated or altered, including with AI. Its examples include a reel narrated with a realistic AI-generated voiceover. Images do not require a label from you, though Meta may label them if its systems detect AI.

You add the label with the "Add AI label" toggle when you create a post, reel or story. There is also an account-wide AI-generated profile label, which matters for AI UGC accounts: Instagram's recommendation guidelines say it tries not to recommend accounts that repeatedly share content prominently featuring a photorealistic AI-generated person, unless the account carries that profile label.

Meta says there may be penalties if you do not label content as required. It may also apply a label itself when it detects AI signals, including industry-standard metadata from third-party tools.

YouTube

YouTube's AI disclosure policy requires disclosure of AI-generated or meaningfully altered content that seems realistic. It lists three cases: content that makes a real person appear to say or do something they didn't, alters footage of a real event or place, or generates a realistic scene that didn't happen. You disclose during upload: in the Attributes section, under "AI use", choose Yes.

YouTube says disclosing AI content won't limit a video's audience or its eligibility to earn money. Cloning your own voice for voiceovers is on its list of uses that don't need disclosure. YouTube may add a label itself for content made with its own AI tools, content carrying C2PA metadata, or content its systems detect as AI.

Under YouTube's monetization policies, channels using AI personas presented as human experts on health, legal, financial or political topics are not allowed to monetize.

4. What happens if you don't label

Platform	Stated consequence
TikTok	Unlabeled content may be removed, restricted or labeled by TikTok, depending on potential harm
Instagram	Possible penalties; Meta may apply the label; accounts with photorealistic AI people and no profile label may not be recommended
YouTube	Creators who consistently don't disclose may get a label applied manually, content removed or YouTube Partner Program suspension

A label is not a pass, either. TikTok, for example, still prohibits AI content that misleads on matters of public importance or harms individuals, whether or not it is labeled.

A working checklist

Written consent for every real face and voice, covering cloning, ads and withdrawal.
No invented customer experiences from AI presenters, or a clear disclosure that they are not real customers.
Commercial disclosure on promotional posts, using the platform setting plus your own on-screen text.
AI label on for realistic AI people or voices: TikTok's AIGC label, Meta's AI label, YouTube's "AI use" answer.
A profile-level label on Instagram for accounts built around an AI character.
Leave C2PA metadata in place. Platforms read it to apply labels automatically.

In Reachira, creating a character from photos of a real person requires consent that the person's identity may be trained, and a cloned voice needs its own consent. The "AI-generated content" flag is a setting you control on each post and campaign. It is off by default, and Reachira passes your choice to each platform's disclosure field without changing it. Whether a video needs the label is your call, made against the rules above. The AI character format and the acceptable use rules explain the rest.

Sources

Practical notes on short-form video for software. No spam; unsubscribe any time.