AI Influencer Disclosure Rules: What to Label on Meta, TikTok, YouTube, LinkedIn and X

AI influencer disclosure is two labels: one says the content is AI, the other says the post is an ad. A sponsored post from a realistic persona usually needs both. As of October 2026, TikTok and YouTube require the AI label on realistic AI content, Meta requires it on photorealistic AI video and realistic AI audio (not still images), and we found no general AI-label rule on LinkedIn or X. Beyond the platforms, the EU AI Act requires deep fakes to be disclosed from 2 August 2026, ad or not, and New York, the UK ad regulator and India add rules for AI people in ads.

Oakgen is our product; it makes AI images, video and characters. Labelling stays your job, and nothing here says an Oakgen output is compliant on its own.

AI influencer disclosure rules at a glance

Each row links to the page holding the rule; dates are as the source prints them.

A decision flow to run on every post

Each "yes" adds a duty; none cancels an earlier one.

1. Is it realistic? A photoreal face, a realistic voice, a scene that looks real. Then turn on the AI setting on TikTok and YouTube, and on Meta for video or audio. Plainly cartoon or anime personas mostly escape platform AI labels (TikTok exempts "artistic styles, like anime"). Our reading for Meta: a photoreal still of an AI influencer doesn't need Meta's label; the same persona in a Reel with an AI voice does. That is only the platform rule. A still-image ad can still need AI disclosure under New York's law (visual ads), the EU AI Act (deep fakes) or ASCI in India (synthetic influencers in ads).
2. Does it look or sound like a real person? Check consent before labels. TikTok says its guidelines "do not allow AI-generated content that contains the likeness of any private individuals without their consent" (Creator Academy), and its ad policy bans using a public figure's likeness without permission, including "making a false endorsement". A label doesn't fix that. If the persona speaks in a cloned voice, read our ethical voice-cloning guide first.
3. Is it an ad? Paid, gifted, affiliate, or a brand's own persona selling its own product. Then you need an ad disclosure as well as the AI label.
4. Is it political or a social issue? Meta has separate disclosure rules for those ads. Our advice: keep AI personas out of political content unless a lawyer has signed off.
5. Where is the audience? Our reading of how far each rule reaches: EU viewers bring the AI Act, a New York ad brings the synthetic performer law, UK ads fall under the ASA and Indian ads under ASCI.

Platform rules: Meta, TikTok, YouTube, LinkedIn and X

Meta (Instagram, Facebook, Threads)

Meta's misinformation policy says: "We require people to disclose, using our AI-disclosure tool, whenever they post organic content with photorealistic video or realistic-sounding audio that was digitally created or altered, and we may apply penalties if they fail to do so." Its help page (accessed 22 Oct 2026) says "This requirement does not apply to images".

Meta uses two names for its label. Its labelling page (updated February 19, 2025) says labels "are now displayed as 'AI Info'", while the current help page says detected content is "labeled AI content". The help page also says a creator who mostly posts AI content "can add an 'AI creator' label to their profile", which can appear next to the account's posts.

TikTok AI label

TikTok's Community Guidelines, in force since 24 September 2026, say "we require creators to label AI-generated or significantly edited content that shows realistic-looking scenes or people" and "Unlabeled content may be removed, restricted, or labeled by our team, depending on the harm it could cause." The setting is in the checklist below. The Creator Academy adds: "you cannot edit or remove the label after posting your video at this time."

YouTube

"We require creators to disclose AI-generated or altered content that seems realistic." That covers 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" or "Creates music that's the main focus of the video" (YouTube help). Creators who "consistently choose not to disclose" risk a label, removal or Partner Program suspension. YouTube also labels some content itself, and labels on "content containing C2PA metadata... can not be adjusted."

LinkedIn

LinkedIn requires no AI label that we found. It shows "the C2PA icon" on media signed with Content Credentials, ads included (LinkedIn help), and recommends saying so when you relied heavily on AI.

X

X's Authenticity policy bans "inauthentic media... that may result in widespread confusion on public issues, impact public safety, or cause serious harm", including "media depicting a real person that has been fabricated or simulated", and bans "AI-generated profile photos... for the purpose of deceiving others". Our reading: an openly AI persona isn't what the fake-persona rule targets; one passing itself off as human is much closer to it.

Ads and sponsorship: the AI label is not the ad label

The AI setting and the ad label are separate switches. Turning one on does nothing for the other.

FTC. When a connection between endorser and seller "might materially affect the weight or credibility of the endorsement, and that connection is not reasonably expected by the audience, such connection must be disclosed clearly and conspicuously" (16 CFR 255.5). The FTC's Disclosures 101 says "ad" and "sponsored" work, warns "Don't assume that a platform's disclosure tool is good enough", and for video says "the disclosure should be in the video and not just in the description uploaded with the video."

UK ASA. Ads must be "obviously identifiable as such". The ASA and CMA "prefer labels that explicitly call out the content as 'advertising' e.g., 'Ad'", a disclosure only after "See more" is not acceptable, and the ASA's 2026 research found "Paid partnership" "may not be enough on its own" (ASA).

Platform tools. YouTube's paid promotion button "adds a disclosure label that appears at the beginning of your video" (YouTube help). On X, sponsored organic posts "must have the 'Paid Partnership' disclosure on", set in Content Disclosure settings (X policy).

Paid ads containing AI. TikTok's ads policy (updated April 2026) allows AI media if you "Apply the AIGC label" or add "a clear disclaimer, caption, watermark, or sticker of your own"; otherwise the ad "will be rejected or restricted". In Ads Manager you "check the toggle This ad contains AI-generated content"; Spark Ads follow the organic Community Guidelines label instead, and after you submit "you cannot change the toggle or unselect it unless you change the video or image" (TikTok ad disclaimers). Meta requires advertiser disclosure for social issue, election and political ads. For other ads, its help page describes Meta adding "AI info" itself under "About this ad". We found no general self-disclosure duty for ordinary Meta ads; the FTC and ASA duties still apply.

Rules written for AI personas

The FTC treats a persona as an endorser. The endorser "could be or appear to be an individual, group, or institution." The FTC rejected a reading limited to virtual influencers: the language is "intended to also encompass the writers of fake reviews and non-existent entities that purport to give endorsements" (Federal Register, Jul 26, 2023). Our reading: a sponsored AI persona post needs the same disclosure as a human's.

Fake reviews are banned outright. The FTC's rule makes it unlawful for a business to create a review or testimonial that misrepresents "That the reviewer or testimonialist exists" or that they used the product (16 CFR 465.2), and the preamble says "AI-generated reviews are covered by the final rule." Our AI UGC disclosure checklist covers testimonials in more depth.

New York names synthetic performers. Anyone producing an ad for a commercial purpose who has actual knowledge "shall conspicuously disclose in such advertisement that a synthetic performer is in such advertisement". A synthetic performer is an AI or algorithm-made asset meant to look like a human performer "who is not recognizable as any identifiable natural performer". Audio ads, translation-only AI, and ads for films, TV and games whose synthetic performer use is consistent with the work itself are exempt. Penalties are $1,000, then $5,000 for later violations (S8420-A). The Governor's office announced the law in effect in June 2026. Our reading: an invented AI influencer in an ad fits the definition closely.

The UK ASA typically defines an influencer as "any human, animal or virtually produced persona" active on social media. It has no AI-specific rules, but asks "Is the audience likely to be misled if the use of AI is not disclosed?", says disclosing "that an influencer is AI-generated" could help, and warns "disclosure alone is very unlikely to mitigate the harm caused by a fundamentally misleading message" (ASA, May 29, 2025). In the ShroomIQ ruling (April 22, 2026), AI-generated "experts" on a supplement site breached the Code despite a footer saying the people shown "are AI-generated representations".

India's ASCI wants two disclosures. Its influencer guidelines (1.5) say "A virtual influencer must additionally disclose to consumers that they are not interacting with a real human being. This disclosure must be upfront and prominent." Its AI labelling guidelines, dated September 17, 2026, make labelling mandatory for "synthetically generated influencers and ambassadors" and suggest labels such as "Audio/Video created using AI".

The EU AI Act. Article 50(4): deployers of AI that "generates or manipulates image, audio or video content constituting a deep fake, shall disclose that the content has been artificially generated or manipulated." A deployer is anyone using an AI system under its authority, "except where the AI system is used in the course of a personal non-professional activity". A deep fake is content that "resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful". For "evidently artistic, creative, satirical, fictional or analogous" work, the duty shrinks to a disclosure "that does not hamper the display or enjoyment of the work." It must come "at the latest at the time of the first interaction or exposure", from 2 August 2026; the 2026 Digital Omnibus did not move this duty.

Fines for breaching Article 50 reach €15 million or 3% of worldwide annual turnover, whichever is higher. For SMEs, including start-ups, it is whichever is lower (Article 99(6)), and the Omnibus extends the lower cap to small mid-caps (new Article 99(6a)). Our reading: most AI-influencer operators are likely SMEs. Whether a wholly invented AI influencer is a "deep fake" turns on the "resembles existing persons" wording, and we don't take a view. With an EU audience, the cautious path is to disclose at first exposure. For the marking and metadata side, see our content provenance guide.

Pre-publish labelling checklist

Run this before every post. It only lists platform steps we could confirm on official pages in October 2026.

Mistakes that cause trouble

Disclosing only in a hashtag, a comment or the bio. The FTC says "Don't mix your disclosure into a group of hashtags or links", and the UK ASA rejects disclosures that appear only after "See more". ASCI says disclosures in a bio or profile are "likely to be missed", so per-post rules still apply.

AI testimonials and reviews. Our reading: the riskiest persona post is one that sounds like a customer. The FTC rule covers reviewers who don't exist, ASCI's AI guidelines (1B) bar exaggerating product results through claims or visuals, and ShroomIQ shows a footer disclaimer didn't rescue invented experts. Our AI UGC disclosure checklist has safer script patterns.

Keep a record of how each post was made

Our advice, not a rule we found: keep a per-post log of tools and model versions, prompts and references (with consent for any real person), edits, the labels you applied with a screenshot of the live post, and who paid and approved. The evidence log template in our AI UGC checklist is a good starting sheet.

Where Oakgen fits, and where it doesn't

Oakgen's AI Influencers studio lets you "Create someone once. Bring their face and voice into any image, video or voiceover." Our drift test shows how to check the persona stays consistent. Disclosure is still yours: the label, the platform toggles and the ad disclosure, every post. Plans are on the pricing page.

For a one-off cartoon mascot you don't need a persona tool. For the business side, see how AI influencers make money, what brands need to know about AI avatars and UGC-style ads without paying influencers.

Frequently asked questions

Do I have to label an AI influencer on Instagram?

As of October 2026, Meta's platform rule requires a label on photorealistic AI video and realistic AI audio, but not on still images, though Meta may label detected AI images itself. Laws can still require AI disclosure on a still-image ad, for example in New York, the EU or under India's ASCI, and a sponsored post also needs an ad disclosure.

Is a made-with-AI hashtag enough?

Usually not on its own. TikTok and YouTube have their own AI settings, and the FTC warns against mixing a disclosure into a group of hashtags. Use the platform setting plus plain words in the caption or on screen.

Does a virtual influencer have to say a post is sponsored?

Our reading is yes. The FTC's endorsement guides cover endorsers who only appear to be a person, the UK ASA typically defines influencers to include virtually produced personas, and India's ASCI makes virtual influencers disclose they are not human on top of the ad label.

Is my fictional AI influencer a deepfake under the EU AI Act?

The Act defines a deep fake as content that resembles existing persons, objects, places, entities or events and would falsely appear authentic. Whether a fully invented persona meets that test is a legal question we don't answer; ask a lawyer in the market you target.

Can I use AI-generated testimonials in ads?

Not presented as real customers. The FTC's rule on fake reviews bans testimonials that misrepresent that the reviewer exists or used the product, and the FTC says AI-generated reviews are covered.

Can I remove TikTok's AI label after posting?

No. TikTok's Creator Academy says you cannot edit or remove the label after posting at this time, so decide before you publish.

Does LinkedIn require an AI label?

As of October 2026 we found no mandatory AI label in LinkedIn's own rules. It shows a C2PA icon on files that carry Content Credentials and recommends telling people if you relied heavily on AI. That is a platform rule only; legal duties such as ad disclosure still apply.