AI presenters make it possible to produce a creator-style ad without a creator. They also make one old mistake much easier to make at scale: putting a customer's experience in the mouth of someone who never had it.
This guide sets out what the US Federal Trade Commission's rule on reviews and testimonials says, why it matters so much for AI presenters in particular, and how to write ads that keep the format without the risk. It is general information to help you reduce risk, not legal advice. For your category and your markets, talk to a lawyer.
The short version
An AI presenter can demonstrate the product, describe it, and present the brand's own claims, as long as those claims are true and supported. An AI presenter cannot speak as a customer, report an experience of using the product, or describe results nobody had.
Everything below is detail on those two sentences.
What the rule says
The FTC's Trade Regulation Rule on the Use of Consumer Reviews and Testimonials was announced in August 2024 and took effect on 21 October 2024. Among other things, it prohibits businesses from writing, creating, buying or disseminating consumer reviews or testimonials that misrepresent that they are by someone who does not exist, or by someone who did not have actual experience with the business or its products, or that misrepresent the experience of the person giving them. In announcing the rule, the FTC said explicitly that this covers fake reviews generated by AI.
The rule also covers buying reviews, undisclosed reviews by a company's own staff or relatives, review sites that present themselves as independent while being controlled by the company, suppressing negative reviews, and selling or buying fake indicators of social influence such as followers or views.
Because it is a trade regulation rule, the FTC can seek civil penalties for knowing violations, on top of its other remedies.
It sits alongside the FTC's Guides Concerning the Use of Endorsements and Testimonials in Advertising, last revised in 2023, which say that endorsements must reflect the honest opinions, findings, beliefs or experience of the endorser, and that an advertiser cannot make a claim through an endorser that it could not make directly.
Why AI presenters are a special case
A human creator who has used your product can, honestly, say so. An AI presenter cannot, because there is nobody there to have used it. So any line that puts a first-hand experience in an AI presenter's mouth is, by construction, a testimonial from someone who does not exist, or who had no experience of the product. That is exactly what the rule prohibits.
The same applies to actors reading a script. An actor who is presented as a real customer, speaking about results they did not have, is misrepresenting the experience behind the words.
The practical rule is simple: the presenter demonstrates and describes; they never testify.
Lines that cross it, and honest rewrites
The intent of most of these lines can be kept. What changes is the frame: from "I experienced this" to "here is what it does."
| Crosses the line | Keeps the intent, honestly |
|---|---|
| "I've used this every morning for three months." | "Here's what it does every morning." |
| "My skin cleared up in a week." | Show the product in use, and say only what the brand can substantiate. |
| "This saved my business." | "Built for small shops that ship their own orders." |
| "My kids love it." | "Made for families with young kids." |
| "Five stars, I'd buy it again." | A real rating the brand holds, stated accurately: "Rated 4.7 on our store." |
| "Everyone's switching to this." | Drop it, or use a real, dated figure the brand can show. |
In notes, captions and on-screen text, call the person a presenter, not a customer.
What an AI presenter can say
- Demonstrations: "Watch what happens when you pour it at seven and open it at seven."
- Descriptions: "It's double-walled steel, and the lid locks shut." "It's built to keep coffee hot all morning."
- The brand's own supported claims: a specification, a price, a guarantee, a certification the product really holds.
- Real reviews and ratings, accurately represented: quoted as reviews from real customers, with the source, and without cherry-picking in a way that misleads.
- The brand's voice, clearly introduced: "At Halden we built…" works when the presenter is plainly speaking for the brand. It does not work for a stranger presented as the founder.
Claims that need more than honesty
Some claims are not about who says them but about what they say. These need evidence whoever delivers them:
- Health, body and weight claims. Curing, treating or preventing anything, weight lost, skin changed. The FTC expects health claims to be backed by competent and reliable scientific evidence.
- Money claims. Income earned, money saved, returns. Platforms restrict these heavily as well.
- Before-and-after. A before-and-after that implies typical results is a results claim. A small "results may vary" line does not fix a claim you cannot support. If you use the format at all, keep it clearly illustrative and label it as such.
- Scarcity and urgency. "Only three left" and "ends tonight" only for a real offer with a real deadline.
- Comparisons. Naming or pointing at a competitor needs a basis you can show.
- People. No real, identifiable person's face or voice without their consent, and every presenter an adult.
Labelling AI presenters
Disclosure is a separate duty from honesty. A label does not make a fake testimonial acceptable, and an honest demonstration may still need a label.
- TikTok requires creators to label AI-generated content that contains realistic images, audio or video, and can apply an AI-generated label automatically to content carrying Content Credentials.
- Meta labels AI-generated content on Facebook and Instagram and asks people to disclose photorealistic video or realistic-sounding audio that was digitally created or altered.
- YouTube requires creators to disclose realistic content that was altered or synthetically generated.
- In the EU, the AI Act's transparency obligations apply from 2 August 2026, including a duty to disclose that realistic generated or manipulated content was made with AI.
Platform policies change, and ads can have their own rules on top of the ones for posts, so check each platform's current advertising policy before a campaign goes live.
A checklist before you publish
- No line in the first person about using the product or its results.
- Every number, rating, review and award is in the brand's own data, current and supported.
- Health, body and money claims are substantiated, or gone.
- Offers and deadlines are real.
- No competitor named without a basis.
- No real person's likeness or voice without consent; every presenter an adult.
- The AI label is on wherever the platform expects it.
How Vessero handles it
Every person in our own examples and samples is AI-generated, and our Trust page says so. Our free hook generator and UGC script generator write to the rules above today: they never write a line in a customer's voice, never use a number you did not give them, and never fill the reason-to-believe beat with a claim you cannot back up.
Three things are being built and are marked as coming across the site until they ship: Cast, which puts AI presenters into your campaigns and labels every one of them as AI; checks on each ad against the platform policies it will run under, before export; and Content Credentials written into what you export.