How Brands Should Disclose AI Influencers

How Brands Should Disclose Synthetic People

A brand should clearly disclose a synthetic person when the artificial identity, use of a real person's likeness, sponsorship, claimed experience, or realistic production could change how a reasonable audience interprets the message.

One generic "AI used" notice is often too vague. A campaign may need to explain that the character is fictional, that the image or voice is synthetic, that the post is advertising, and that any product experience comes from a real authorized source rather than the avatar.

The correct disclosure depends on the market, medium, platform, and facts. This guide provides a review framework, not a universal safe harbor or legal opinion.

First identify what the audience may believe

"AI influencer" can describe several different arrangements. A wholly fictional character may promote a product under a consistent invented identity. A generated avatar may present words written by a brand employee. A synthetic double may resemble a real actor. A cloned voice may speak new lines. A virtual character may repeat a real customer's testimonial. A generated person may imply first-hand use that never happened.

The review starts with the likely audience impression, not the production team's label for the asset.

The table deliberately separates identity from claim. A fictional character can deliver a false product statement. A real person can deliver a properly substantiated advertisement. Synthetic production is one dimension of the review.

Separate four disclosures

The first question is whether the character or performance is synthetic. The second is whether a real person's likeness or voice is being used. The third is whether the message is sponsored or controlled by a marketer. The fourth is whether the content presents a testimonial or first-hand experience.

These facts should not be collapsed.

A suitable disclosure might say that a spokesperson is a fictional, AI-generated character created for the brand. If the post is paid influencer advertising, the material connection also needs a clear disclosure. If the character presents a real customer's statement, the brand should identify the representation without implying that the avatar personally used the product.

The exact wording should be tested in the final layout and medium. The audience should understand the important fact before relying on the face, voice, demonstration, or endorsement.

United States rules focus on deception and testimonials

The United States Federal Trade Commission does not impose a blanket prohibition on virtual influencers. In its current Consumer Reviews and Testimonials Rule Q&A, FTC staff says an AI stock avatar is not itself a consumer review, but its use may be a testimonial. A fake or false underlying testimonial can violate the rule, and an avatar can still be deceptive under the FTC Act.

The same Q&A addresses a celebrity avatar used without permission. If reasonable consumers would think the celebrity gave the favorable testimonial, the use can violate the rule.

FTC Disclosures 101 guidance says a material connection should be disclosed clearly and conspicuously with the endorsement. A synthetic-identity notice does not replace that connection disclosure.

These authorities do not answer every question about publicity rights, copyright, biometric data, labor, privacy, or state law. A real-person likeness or voice requires specific legal review for the places where the campaign will run.

Platform disclosure is an additional layer

YouTube requires creators to disclose meaningfully altered or synthetically generated content when it seems realistic. Covered examples include making a real person appear to say something they did not say and generating a realistic scene that did not occur. The creator selects the altered-content setting, and YouTube displays information to viewers.

That platform control does not replace the words needed in the content. It also does not replace a paid-promotion declaration, an endorsement disclosure, or consent.

flowchart TD
    A["Synthetic person in brand content"] --> B{"Could viewers think a real person exists, participated, or spoke?"}
    B -->|Yes| C["Disclose synthetic identity or performance"]
    B -->|No| D["Confirm fictional nature is obvious"]
    C --> E{"Is a real likeness or voice involved?"}
    D --> E
    E -->|Yes| F["Verify specific authorization and market law"]
    E -->|No| G["Continue claim review"]
    F --> G
    G --> H{"Endorsement, testimonial, or sponsorship?"}
    H -->|Yes| I["Substantiate experience and disclose connection"]
    H -->|No| J["Review ordinary advertising claims"]
    I --> K["Apply platform and jurisdiction rules"]
    J --> K

European Union transparency changes the market review

The European Union AI Act applies generally from August 2, 2026. Article 50 includes transparency duties for certain AI-generated or manipulated image, audio, and video content. It defines deep fakes around content that resembles existing persons, objects, places, entities, or events and could falsely appear authentic or truthful.

The regulation includes a more limited treatment for evidently artistic, fictional, satirical, creative, or analogous works, but still calls for an appropriate disclosure of the generated or manipulated nature. The exact duty depends on role and facts.

A campaign available in the European Union should not rely on a United States-only influencer analysis. It needs a current Article 50 review, applicable member-state context, and any sector rules.

United Kingdom advertising adds another evidence question

The UK advertising codes require testimonials and endorsements to be genuine unless they are obviously fictitious, with documentary evidence supporting them. The ASA and CAP guidance should be checked for the current rule and medium.

A photorealistic generated person praising a product creates a direct question: will the audience understand that the speaker is fictional, or will it infer a genuine user's experience? Clear creative intent does not cure an unsupported product claim.

Put the disclosure where the impression occurs

For an image, place the disclosure in or immediately beside the image and preserve it in repostable versions when practical. For video, disclose early enough that the audience does not first absorb a false impression, and repeat the information when the synthetic person appears after a long gap. For audio, use spoken disclosure rather than relying only on show notes. For a persistent virtual influencer, make the synthetic identity clear in the profile and in individual commercial posts where needed.

Readable language matters. "Fictional AI-generated spokesperson for Brand X" says more than "enhanced media." "Synthetic recreation of Jane Doe's voice, used with permission" says more than "AI-assisted."

Test the final version on someone who did not attend production. Ask who they think the person is, whether they believe the person used the product, and whether they understand who paid for the message.

Keep a campaign record

Preserve the character design, model and vendor, prompts or instructions needed to explain the workflow, source assets, likeness and voice permissions, contracts, claim substantiation, underlying testimonials, approved wording, platform settings, market list, legal review, publication URLs, and corrections.

[[How Creators Can Build an Authenticity Record]] provides the broader asset workflow. A brand campaign should add approval ownership, market-specific decisions, monitoring, and a takedown or correction path.

The safest operational rule is simple: do not let the synthetic person claim an identity, relationship, participation, or experience that the evidence and disclosure cannot support.

This guide reflects FTC, YouTube, European Union, and United Kingdom primary guidance reviewed on July 27, 2026. It is general information, not legal advice. AI assistance was used for research organization, drafting, and validation. Publication remains unauthorized.