AI disclosure on Reels: FTC, New York and platform rules
A properly labeled AI Reel costs you nothing. A missing label can cost a lot: platform penalties, an embarrassed client, lost trust. The good news: the rule comes down to two steps.
The simple rule
Whenever an AI-generated face or voice appears in a Reel:
- put a clear on-screen disclosure on the video for its full length, such as "AI-generated";
- turn on the platform's AI label when you post.
If the post is an ad or a paid partnership, disclose that too ("Ad", "Paid partnership"). Kelpo adds the on-screen disclosure on request, in the chat or with one click when you download.
Federal law: no "AI label" rule, but no deception either
There's no federal rule that requires an "AI" label as such. What applies is the FTC Act's ban on deceptive advertising:
- The Endorsement Guides (16 CFR Part 255, updated in 2023) treat anyone who appears to be an individual, virtual influencers included, as an endorser. Endorsements must reflect honest opinions and real experience, and material connections must be disclosed clearly and conspicuously. A cloned owner may only say what the owner actually believes and can back up.
- The Rule on Consumer Reviews and Testimonials (16 CFR Part 465, in effect since October 21, 2024) bans fake reviews and testimonials, and the FTC names AI-generated ones explicitly. Knowing violations can bring civil penalties.
New York: synthetic performers in ads
New York's synthetic performer disclosure law (S.8420-A, signed December 11, 2025, in effect since June 2026) requires ads that use a "synthetic performer" to say so conspicuously. A synthetic performer is a digitally created human-looking performer who isn't any identifiable real person: a stock AI actor, in other words. Penalties: $1,000 for a first violation, $5,000 for each one after that.
The law doesn't cover replicas of a real, recognizable person, like your client's cloned owner. Those fall under state right of publicity laws, which require the person's consent: for example California Civil Code §3344 and New York Civil Rights Law §§50-51.
On the platforms
| Platform | What's required | Where to click |
|---|---|---|
| TikTok | Label required for AI-generated content with realistic images, audio or video | "More options," then "AI-generated content" |
| Instagram, Facebook, Threads | Label required for photorealistic video or realistic audio made or altered with AI; Meta may apply penalties if you skip it | "Advanced settings," then the AI label |
| YouTube Shorts | Disclosure required for realistic content, including a real person shown saying something they didn't say: a cloned owner counts, even with consent | "Altered or synthetic content," then "Yes" |
| No mandatory label | The on-screen disclosure is enough |
The label won't appear on its own. Kelpo writes an "AI-generated with Kelpo" tag into the MP4's metadata, but it isn't a C2PA Content Credential, so platforms may not detect it. Turn the label on every time you post.
Where the disclosure goes
Near the top of the frame, clear of the platform's buttons and caption, white on a dark background, big enough to read on a phone without zooming. On screen from the first second to the last.
Three hard rules
- An AI actor never poses as a customer. The actor presents the offer; they don't give a testimonial about it.
- No replica without a signed release from the person: term, permitted uses, revocable anytime.
- No minors, and no health content that promises results.
If your client also targets the EU
Under the EU AI Act (Article 50), anyone publishing a deepfake must disclose that it's AI-generated, since August 2, 2026. A cloned owner is the textbook case. Keep the on-screen disclosure on every version.
One more line on your proposal
The disclosure isn't something to hide. It's a service: put it on the proposal, handle it for your client, and they never have to think about it. See how to sell AI Reels.
Sources, accessed October 12, 2026: 16 CFR 255, 16 CFR 465, NY S8420A, TikTok, Meta, YouTube, LinkedIn. This article isn't legal advice.