Notes
AI-Generated Ads and the FTC: What DTC Brands Need to Know in 2026
Published
There is no FTC rule against AI-generated advertising. There is a rule against fabricated endorsements, and it does not care whether the fabrication was done with a camera or a model.
The rule people expect, and the rule that exists
Brands usually come to this looking for a rule about AI. There isn't one, in the sense of a regulation that says AI-generated ads are permitted or forbidden. What there is instead is a rule about fabricated endorsements, and it lands squarely on the most common way AI video gets used.
In August 2024 the FTC finalised a rule banning fake reviews and testimonials, including those attributed to someone "who does not exist" — see the FTC's August 2024 final rule. It carries civil penalties. It was written with review farms and AI-generated review text in mind, and it applies just as cleanly to a synthetic person on camera describing a result they never had.
Why this catches AI UGC specifically
The dominant format in this category is a person talking to camera about their own experience with a product. That format *is* a testimonial. Generate the person and you have produced a testimonial from someone who does not exist, which is the exact thing the rule names.
That's not a technicality you can write around with a disclosure. A label saying the video is AI-generated tells the viewer how it was made; it does not make the endorsement real. The underlying claim — this person used it, this happened to them — is still false.
The distinction that keeps you clear
An AI performer can be a brand spokesperson. A spokesperson presents the product, makes claims the brand can substantiate, and doesn't purport to be an independent customer sharing their own results. That's an ad with a presenter in it, which advertising has done for a century.
An AI performer cannot be a customer. The moment the script says "I tried this and", you've crossed from presentation into endorsement, and the endorser doesn't exist.
The full version of that argument, with script-level examples, is in AI spokesperson vs fake testimonial.
The Endorsement Guides still apply on top
the FTC's Endorsement Guides require that an endorsement reflect honest opinions and that claims be substantiated. That obligation attaches to the advertiser regardless of who or what delivers the line — so a synthetic presenter making an unsubstantiated efficacy claim is the brand's problem, not the tool's.
What this means in practice
- Don't script first-person experience claims for a synthetic performer.
- Don't invent a name, a backstory, or a purchase for them.
- Do disclose AI generation per platform policy — see Meta's labelling and TikTok's.
- Do keep substantiation for every product claim, exactly as you would for a live-action ad.
- Do keep a record of what was generated and when, in case anyone asks later.
This isn't legal advice. It's an operator's account of rules we work under every day. If you're making a call with real money or real exposure behind it, get it reviewed by someone qualified.
It's why our terms say we won't produce fake testimonials — it's a term of the agreement rather than a marketing position.
Read next
Start here
One angle, two finished videos, three business days. $50, one-off, and you own every file.