Two disclosures, and they answer different questions. The sponsorship disclosure tells the audience money changed hands, and that's required under the FTC's endorsement guides. The AI disclosure tells them the content is synthetic, and there's no federal rule specifically requiring it — though New York's synthetic performer law does, and every major platform does. Neither disclosure satisfies the other.
The common mistake isn't failing to disclose. It's disclosing once and assuming that covered both.
Why are these two different obligations?
Because they protect against two different things.
A sponsorship disclosure exists because a viewer weighs a recommendation differently when they know it was paid for. An AI disclosure exists because a viewer weighs a depiction differently when they know it isn't real.
A synthetic presenter in a paid post creates both problems at once. Telling the audience it's an ad doesn't tell them the person isn't real. Telling them the person isn't real doesn't tell them it's an ad.
One says money changed hands. The other says the person didn't exist. A viewer needs both.
What does the FTC actually require?
Material connection disclosure — and the framework reaches synthetic personas, which is the part most creators don't realise.
The binding document is 16 CFR Part 255, the Guides Concerning Use of Endorsements and Testimonials in Advertising. These are administrative interpretations of Section 5 of the FTC Act, and they were revised in 2023 — the first substantive update since 2009.
The 2023 revision matters here for one specific reason. It redefined an endorser as a party who may appear to be an individual, group or institution. Law firms reading that language conclude it encompasses virtual influencers, writers of fake reviews, and non-existent entities purporting to give endorsements.
So a synthetic persona endorsing a product in a paid post is an endorser. The material connection between the brand and that endorsement has to be disclosed exactly as it would if a real person delivered it.
What does "clear and conspicuous" mean?
It has a definition, and it's stricter than most disclosures manage.
Difficult to miss, and easily understandable by ordinary consumers.
That's the standard. Two parts, and most failed disclosures fail the first rather than the second.
A hashtag buried among fifteen others is not difficult to miss. Neither is a disclosure in a pinned comment, mid-description, or behind a "more" tap. The test looks at how an ordinary viewer actually perceives it rather than whether the information was technically present.
Where this falls short. "Difficult to miss" is a standard rather than a specification, and reasonable people place disclosures differently. What's clear is the direction: if a viewer could plausibly consume the whole piece without seeing it, it isn't conspicuous.
Is there a federal rule requiring AI disclosure in ads?
No, and this is worth being precise about because a lot of writing implies otherwise.
There is no US federal rule that specifically requires disclosure of AI-generated content in advertising. What exists is Section 5 of the FTC Act prohibiting deceptive acts and practices, and the endorsement guides applying it. Undisclosed synthetic content becomes an FTC problem when it creates a false impression that would matter to a consumer's decision — not automatically, and not as a labelling requirement.
That's a meaningful distinction. The obligation is to avoid deception, not to affix a label. Whether a given piece deceives depends on what it depicts and what the audience would reasonably assume.
One thing that follows from it. A fictional dramatization has never been an endorsement. The endorsement guides' own long-standing example is two unnamed actors in a supermarket praising a detergent — an obvious fictional dramatization of a real-life situation, and not an endorsement. So narrative advertising and testimonial advertising are different categories, and the second one is where the exposure sits.
Which law does require it?
One, and it's a state law rather than a federal one.
New York's S.8420-A / A.8887-B requires commercial advertisers to disclose conspicuously to consumers when a synthetic performer is used in a visual or audiovisual advertisement.
The definition is specific: a digitally created asset created, reproduced or modified by computer, using generative artificial intelligence or a software algorithm, that is intended to create the impression that the asset is engaging in an audiovisual or visual performance.
This is currently the only law in the United States that specifically requires a disclosure for the use of AI-generated content in advertisements. Where it applies, it goes beyond the general deception standard — it requires a disclosure whether or not anyone would be deceived.
Its application is limited, though, by the type of advertisement, the type of generated content, and the advertiser's intent. It isn't a blanket labelling rule.
Who is actually liable?
More parties than people expect, and the answer isn't "whoever posted it."
Advertisers bear responsibility for endorsements made on their behalf. Endorsers bear their own responsibility for what they say. And intermediaries — agencies, review brokers, reputation management companies and similar — may be liable for their role in creating or disseminating endorsements they know or should know are deceptive, or that fail to disclose unexpected material connections.
The guides recommend that both advertisers and intermediaries maintain reasonable programmes to train and monitor the endorsers they pay or direct.
For a creator that has a practical edge. A brand telling you the disclosure is handled doesn't discharge your own position, and a contract allocating liability between you doesn't bind a regulator.
How do you actually do it?
Six steps, and the first two do most of the work.
Disclose the commercial relationship first
Difficult to miss and easily understandable. Not a buried hashtag, not a pinned comment, not behind a "more" tap. This is the obligation with the clearest legal basis.
Disclose the AI separately
A separate statement, because it answers a different question. Neither disclosure implies the other, and combining them into one label tends to obscure both.
Set the platform toggle as well
Every major platform runs its own AI disclosure system and one platform's label doesn't travel to another. The toggles are separate from anything you write in the post.
Check whether a state law applies
New York's synthetic performer provision is currently the only US law specifically requiring it, and it reaches advertisers rather than platforms. Others are being introduced.
Don't rely on your generation tool
Most add nothing. Hexcoded states plainly that it adds no AI label and no watermark, and that labelling where you publish is the publisher's obligation. Its acceptable use policy is blunter: the platform provides the tool, not clearance of your campaigns.
Keep the brief and the approval
If a brand directed the synthetic element, that's worth having in writing. It doesn't shift liability, and it's the first thing anyone will ask for.
What about a licensed real person?
Different position again, and it's the case where the paperwork is cleanest.
If the face is a human creator who licensed their likeness, a real person did consent to appear. The endorsement is still an endorsement and the material connection still needs disclosing — but you aren't presenting a non-existent person as real, which is the deception risk the synthetic case carries.
The AI disclosure question doesn't disappear, because the content is still synthetically generated. What changes is that the depicted person exists, agreed to this, and is being paid for it.
On Hexcoded that's documented rather than asserted: human creators pass a liveness check and a face-match before their likeness can be published, their consent is a timestamped record, and they earn on every render. Whether that needs disclosing to your audience is a separate question from whether it's cleared.
This describes US federal and state requirements current as of the publication date. The FTC endorsement guides are administrative interpretations of Section 5 of the FTC Act and their application is fact-dependent. State law varies and is changing. Platform requirements change frequently. Nothing here is legal advice.
- Two disclosures, two reasons. One says money changed hands, the other says the person isn't real
- The FTC's endorsement guides reach synthetic personas. The 2023 revision defines an endorser as anyone who appears to be an individual
- "Clear and conspicuous" means difficult to miss and easily understandable. Buried hashtags fail it
- There is no federal rule specifically requiring AI disclosure in advertising. The obligation is to avoid deception
- New York's synthetic performer law is currently the only US law that does require it
- An obvious fictional dramatization has never been an endorsement. Narrative and testimonial are different categories
- Intermediaries can be liable too, and a contract between you and a brand doesn't bind a regulator
- Your generation tool probably labels nothing. Hexcoded doesn't, and says so
Treat it as two separate obligations. The sponsorship disclosure is required under the FTC's endorsement guides. The AI disclosure has no specific federal requirement, but platform policies require it, New York law requires it for synthetic performers in advertisements, and undisclosed synthetic content can be deceptive under Section 5.
Yes. The 2023 revision defines an endorser as a party who may appear to be an individual, group or institution, which law firms read as encompassing virtual influencers and non-existent entities purporting to endorse. A synthetic persona endorsing a product is an endorser.
Difficult to miss — easily noticeable — and easily understandable by ordinary consumers. The test is how an ordinary viewer actually perceives the disclosure, so a buried hashtag, a pinned comment, or a disclosure behind a "more" tap all fail it.
No. There's no US federal rule specifically requiring it. What applies is Section 5 of the FTC Act prohibiting deceptive practices, and the endorsement guides interpreting it. The obligation is to avoid deception rather than to affix a label.
New York's S.8420-A / A.8887-B, which requires advertisers to disclose conspicuously when a synthetic performer appears in a visual or audiovisual advertisement. It's currently the only US law specifically requiring AI disclosure in ads, and where it applies it goes beyond the general deception standard.
Not necessarily. The endorsement guides' own long-standing example is two unnamed actors in a supermarket praising a detergent — an obvious fictional dramatization, and not an endorsement. Narrative advertising and testimonial advertising are different categories.
Cleared to cast, yours to disclose
Human creators on Hexcoded pass a liveness check and face-match before publication and earn on every render. The disclosures on your published post stay with you — Hexcoded adds no label and no watermark, and says so plainly.
Read the disclosure positionMore on workflow, compliance and delivery in Production.