AI influencers sit where two sets of rules meet: rules about advertising and endorsements, and newer rules about AI-generated content. Brands using an AI character need to respect both. This article summarizes the main frameworks in general terms and shares the checklist we follow when we produce an AI brand ambassador.
Not legal advice. This is a general overview to help you ask the right questions. Rules differ by country, change over time and depend on your specific situation. Check the current official texts and consult a lawyer for your case.
Two separate questions
For any AI influencer content, ask:
- Is it advertising? If a brand pays for or controls the content, the audience must be able to tell it's an ad.
- Is it AI-generated? If the content shows a realistic person, voice or scene created by AI, the audience may need to be told.
The two answers are independent. An AI character promoting your product usually needs both disclosures.
United States: the FTC
The Federal Trade Commission's Endorsement Guides require that any material connection between an endorser and a brand be disclosed clearly and conspicuously. The FTC's updated guides explicitly treat virtual influencers as endorsers, so an AI influencer promoting a brand is subject to the same disclosure expectations as a human one.
The FTC has also adopted a rule against fake reviews and testimonials, which covers reviews that misrepresent who wrote them, including reviews generated by AI. In practice this means an AI character should not pose as a real customer describing a personal experience with the product.
- Disclose the brand relationship in a way people will notice, not hidden in hashtags at the end.
- Don't present AI-generated testimonials as real customer reviews.
- Claims made by the character must be truthful and substantiated, as for any ad.
European Union: the AI Act
The EU AI Act introduces transparency obligations for certain AI-generated content. In general terms, content that is artificially generated or manipulated to resemble real people, objects or events (often called deepfakes) must be disclosed as such by those who deploy it, and providers of generative AI systems must make outputs identifiable as AI-generated in a machine-readable way. These transparency duties apply from 2 August 2026. Existing EU consumer protection law on misleading advertising and hidden marketing continues to apply alongside it.
Meta (Facebook and Instagram)
- Meta labels content it detects as AI-generated or significantly AI-edited and asks creators to disclose realistic AI-generated video and audio.
- Paid partnerships should use Meta's branded content tools so the relationship is visible.
- Ads must follow Meta's advertising policies; ads about social issues, elections or politics have additional AI disclosure requirements.
TikTok
- TikTok requires creators to label AI-generated content that shows realistic scenes, people or audio, and provides an AI-generated content label for this.
- TikTok may label content automatically when it detects AI-generation metadata.
- Commercial content should be marked with TikTok's content disclosure settings.
Our disclosure checklist
| Check | What we do |
|---|---|
| AI label | All AI content is labeled using each platform's tools |
| Ad disclosure | Brand relationship disclosed clearly; branded content tools used where available |
| No fake testimonials | The character explains and demonstrates; it doesn't pose as a customer with a personal result |
| No doctor persona | Characters never present as doctors or give medical advice |
| No generated product results | No AI-generated before/after results in cosmetics, weight loss or hair |
| No financial advice | Characters don't give investment or financial advice |
| Claims review | Every claim checked against the client's brief and substantiation |
Where to put disclosures in a video
A disclosure only works if people see it. For short vertical videos that means:
- Use the platform's own tools first: AI-generated content labels and branded content or paid partnership settings.
- Disclose early. If a disclosure is needed in the video itself, put it at the start, not in the last second.
- Keep it readable. On-screen text should be large enough and stay long enough to read on a phone.
- Say it if the video is spoken. In audio-led content, a spoken mention is clearer than small text.
- Don't bury it. A disclosure hidden among many hashtags or behind "more" is easy to miss.
Common disclosure mistakes
- Labeling the ad as an ad, but not the realistic AI presenter as AI.
- Labeling the AI, but presenting the character as an independent fan rather than a brand voice.
- Using a character to describe "my results" with a product it has never used.
- Assuming that a stylized or cartoon character never needs any disclosure; the advertising relationship still has to be clear.
- Relying on an old checklist; platform tools and rules change, so review them regularly.
What to agree with an agency in the contract
- Who is responsible for labeling and for using the platform disclosure tools.
- Who approves claims before publishing, and how claims are substantiated.
- Who owns the character's image, voice and backstory, and the content library.
- Whether the agency may create similar characters for competitors.
- How the account and content are handed over if the contract ends.
Situations and the disclosures they need
| Situation | AI disclosure | Ad disclosure |
|---|---|---|
| Realistic AI character in an organic post about the brand's product | Yes | Yes, the brand relationship must be clear |
| Realistic AI character in a paid ad | Yes | The ad format already identifies it as an ad; follow the platform's rules |
| Cartoon character on the brand's own account | Follow platform rules for AI content | The account clearly belongs to the brand |
| AI character reviewing a product as if it were a customer | Don't do it: it misrepresents who is speaking |
A pre-publishing checklist
- Is the AI content labeled with the platform's tool?
- Is the advertising relationship clear in the first moments of the video or in the platform's branded content setting?
- Does every claim match the brief and have substantiation?
- Does the character avoid acting as a doctor, a financial adviser or a real customer?
- Are there no AI-generated product results?
- Does the content follow the rules of every country where it will run?
Keep records
If a platform, a regulator or a customer ever asks how a piece of content was made and approved, you want to be able to answer quickly. Keep a simple record for each published video and ad:
- The date, platform and account where it was published.
- Whether the AI label and the branded content setting were applied.
- The claims made and where their substantiation is stored.
- Who approved the content and when.
- The version of the character guide used.
Records like this also make it easier to update content when rules change, because you know exactly which videos to review. A shared spreadsheet or a field in your CRM is enough; what matters is that it is filled in every time.
Other markets
Outside the US and EU, advertising rules still apply. In the UK, ads must be obviously identifiable as ads, and influencer marketing guidance from the advertising regulator applies. The UAE and other markets have their own requirements for advertising and influencers. If you run the same character in several countries, check each market's rules before launching there. For the commercial side of running one character across markets, see AI creators vs paid ads.
What this means for brands
Disclosure doesn't kill performance; misleading content does. An AI character that is clearly labeled and gives genuinely useful content can still build an audience. In our research, follower count barely affected views per video; format and hook did. See followers don't matter and the overview in AI influencers for brands.
Frequently asked questions
Do AI influencers have to say they are AI?
Realistic AI content generally needs to be disclosed under platform rules, and in the EU the AI Act adds transparency duties. Labeling all AI content is the safe practice.
Does the FTC apply to virtual influencers?
Yes. The FTC's updated Endorsement Guides explicitly include virtual influencers, so material connections must be disclosed.
Can an AI influencer give product reviews?
It can explain and demonstrate a product, but it should not pose as a real customer describing a personal experience.
When do the EU AI Act transparency rules apply?
The transparency obligations apply from 2 August 2026. Check the official text for the details relevant to you.
Is this legal advice?
No. It is a general overview. Consult a lawyer for your situation.
Get a free niche audit. On the AI Brand Ambassador page we'll review your niche, including which claims and formats are safe to use.