The FTC ad disclosure is separate from the AI label

A paid integration still needs a clear ad disclosure. An AI info tag does not tell the viewer they are being sold to. Layer the disclosures. Do not hide the paid nature behind the synthetic nature.

The Federal Trade Commission's Endorsement Guides (16 CFR Part 255) are about whether a reasonable consumer can tell there is a material connection, and whether the endorsement reflects an honest experience. They are not a platform sparkle policy. Meta's automatic ads AI info can fire on a file that is not an endorsement at all, and it can fail to mention money on a file that is.

Two questions a viewer can still get wrong

Hold up the finished unit and ask both:

Would this person reasonably think nobody paid for this? If a creator, a "customer," or a founder-looking presenter is promoting a product because of a relationship the viewer would not expect — payment, free product, affiliate, employment — the Guides want a clear, conspicuous disclosure of that connection.
Would this person reasonably think a real customer had this experience, or that a real human is on screen, when that is not true? That is a truth-in-advertising question. Fabricated testimonials are their own problem under the Consumer Reviews and Testimonials Rule. A synthetic presenter framed as "one of our buyers" is a fake review with a face. A synthetic presenter framed as a branded actor is advertising.

AI info, AIGC, Made with AI, altered content — those badges answer a third question: was a generative tool involved. A viewer can read "AI info" in About this ad and still believe the person on screen is a real customer who chose the product. They can read "creator labelled as AI-generated" and still miss that the clip is an affiliate placement. The sparkle does not carry the commercial fact.

What the FTC actually requires

The Guides do not impose a general "if you used a model, say so" duty. Using a generator does not, by itself, create a material connection. The connection is the money, the free SKU, the employment, the affiliate cookie.

What the Guides do require, in the language advertisers have been using for years:

Endorsements must reflect honest opinions and actual experience.
Advertisers are liable for misleading claims made through endorsements, and for failing to disclose unexpected material connections.
Disclosures must be hard to miss. A buried #ad in a hashtag pile, a disclosure only on a linked landing page, or a line after the cut are the patterns the staff guidance keeps failing.

The Consumer Reviews rule then bans fake reviews and fake testimonials, including ones a model wrote, on a per-violation civil penalty. Disclosure does not cure a fabricated experience. You cannot #ad your way out of a customer who does not exist.

So the production rule for UGC sold to brands is not "add AI info and ship." It is:

If it is paid or gifted, say it is an ad, on the creative, in language a muted, vertical, three-second viewer can read.
If the person on screen did not use the product, do not frame them as a customer.
If the person on screen does not exist, do not frame them as a person who exists. A branded actor or an obvious avatar is a different sentence from a testimonial.

An AI info tag is not an #ad

Platform AI labels live in three unhelpful places for commercial disclosure:

Behind a menu. Meta ads AI info is often inside About this ad. The FTC's "clear and conspicuous" test is not a three-dot hunt.
After the hook. YouTube's altered-content note in an expanded description is not on frame zero of a paid Short.
On the account, not the file. Instagram's AI creator badge says the profile uses AI. It does not say this Reel is a paid partnership with a named advertiser.

Use the platform's paid-partnership / branded-content toggle when one exists. Also put the words on screen, because the toggle is not always visible in every placement and is not always passed through when the file is reused. "Ad," "Paid partnership with [brand]," spoken and burned in, in the first seconds, is the version that still works when the unit is watched muted in a story.

AI info can sit next to that line. It cannot replace it.

Layer them, don't swap them

A working stack for one file:

Commercial. Paid-partnership control plus on-screen "Ad" or "Paid partnership with …"
Synthetic, when the picture could pass as real. Platform AIGC / AI Info / altered-content, plus a readable line if the law you are under wants a visible deployer label.
Identity, when a living person's likeness is used. Consent and, where it applies, a synthetic-performer disclosure — New York's rule is the named US example for ads with a fully synthetic human reaching that state.

Swap any layer for any other and a viewer can still be wrong about the fact that layer was supposed to carry. The expensive version of this mistake is hiding a paid integration behind "made with AI," as if the synthetic nature were the news and the money were incidental. It is not incidental. It is the reason the Guides exist.

FAQ

Does an AI info label satisfy the FTC Endorsement Guides?

No. AI info tells a viewer the content may have been generated or edited with a generative tool. The Guides are about material connections and truthful endorsements. A paid integration still needs a clear ad disclosure.

If the presenter is obviously AI, do I still need #ad?

Yes, if money, product, or another material connection is why the clip exists. Obvious synthesis answers "is this a real person." It does not answer "is this a commercial."

Can a disclosure fix a fake customer testimonial?

No. The Consumer Reviews and Testimonials Rule bans fabricated reviews and testimonials. Labelling the file as AI, or as an ad, does not make an invented experience true. Do not generate a customer who does not exist.

Where should the ad disclosure go on a short vertical file?

On the creative, early, readable without sound, not only in a menu, a description, or a hashtag stack. Use the platform branded-content toggle as well when you have one. Neither replaces the other.