Supplement and Skincare Claims in AI UGC Ads: The Three Rulebooks That Reject Your Script

Quick answer:a supplement or skincare script gets rejected for one of three reasons, and they live in three different rulebooks. The regulator (in the US, the FTC) asks whether you can substantiate the claim, and it treats a testimonial as your claim. Advertisers "should not make claims through consumer testimonials or expert endorsements that would be deceptive or couldn't be substantiated if the advertiser made them directly." The ad platforms then apply their own claim rules, which are narrower than most summaries say on Meta and vary by market on TikTok. And if the video runs inside TikTok Shop, a third rulebook written specifically for AI content applies. It is the only one that says an AI visual can be a medical claim "even if no explicit text claim is made." The AI actor changes nothing in the first two and is the whole subject of the third.

Rulebook one: the regulator does not care who is speaking

The FTC's Health Products Compliance Guidance(December 2022) is the document behind every "clinically proven" and "lost 12 pounds" line you have seen bounced, and it never mentions the speaker. Its standard is "competent and reliable scientific evidence": "tests, analyses, research, or studies that (1) have been conducted and evaluated in an objective manner by experts in the relevant disease, condition, or function to which the representation relates; and (2) are generally accepted in the profession to yield accurate and reliable results."

Two sentences from the testimonial section decide most UGC scripts. "It's not enough that a testimonial represents the honest opinion or experience of an endorser. Under FTC law, advertisers also must have appropriate scientific evidence to back up the underlying implied claim that the product is effective and will work for buyers as it did for the endorser." A first-person script on an AI actor is judged exactly like one on a paid creator: the claim is yours and the evidence must be yours. Whether that first-person script is a testimonial at all under the FTC's separate fake testimonial rule is the subject of the FTC and AI actors post.

Then the disclaimer every buyer reaches for: "Moreover, attempts to disclaim dramatic results with statements like 'Results not typical' don't cure the deception. Those testimonials should be accompanied by a clear and conspicuous disclosure of the results a typical consumer can actually expect." Example 41 is a weight-loss supplement ad with before-and-after photos and a 16-pounds-in-eight-weeks quote; where the trial average is four pounds over placebo, "The vague disclosure doesn't adequately convey to consumers that their weight loss is likely to be much less."

For supplement brands leaning on the FDA's structure/function category: "Unlike FDA law, FTC law makes no bright-line distinctions between categories of health-related products or claims." "Supports immune health" is not a softer claim in an ad. It is a claim.

Rulebook two: what Meta and TikTok Ads actually reject

Meta

Meta's Health and Wellness standard, read on 4 September 2026 (its change log shows entries for that day, 23 July 2026 and 27 December 2024), does four things for an ecommerce advertiser.

It sets the age gate: "Ads promoting or marketing dietary, health, or weight loss or weight gain products and services must be targeted to people aged at least 18 years or older." Cosmetic products, procedures and surgeries get the same 18+ rule, but "Non-permanent cosmetic and beauty products such as creams, make-up, hair products and digital editing apps" are exempt, which is most of skincare.

It bans one closed class of cure claim: "Claims (including those from health professionals or health organisations) to cure, heal or eliminate any of the following incurable diseases and/or terminal illnesses (exhaustive list): Diabetes, Herpes, Thyroid, Psoriasis, Ebola, Cancer, Autism, Alzheimer's, Parkinson's, Motor Neurone Disease (MND), Human Immunodeficiency Virus (HIV). Note: This does not apply to claims about treating or managing symptoms of these conditions." Acne, eczema, bloating and hair loss are not on it.

Those fall under the timeframe rule: ads can't "Employ clickbait tactics in a health, weight loss or weight gain context, such as sensational language with exaggerated or extreme claims, or promises of specific outcomes within a set time frame without disclaimers or qualifiers." The condition at the end of that sentence is the whole rule. "Clear skin in 7 days" is the violation; "visibly calmer skin in 7 days in our 8-week study" is what the sentence permits, if the study exists.

And it allows more imagery than the folklore says. At 18+ it permits "General cosmetic products, procedures, operations depicting before-and-after transformation." What it bans alongside, for weight loss or weight gain products, is a "Close up on specific body area by pinching fat", and for both weight and cosmetic ads, "statements of inferiority about physical appearance". A before-and-after is fine on Meta. A fat-pinch in a weight-loss ad is not.

TikTok Ads Manager

TikTok splits the same territory across three policies. The global one is Misleading and False Content(April 2026): "Ad content and landing pages must not promise or exaggerate results concerning a product's effect." Its examples read like a skincare brief: a lotion promising "Get slim legs right away.", "Claims of cures for incurable diseases", and "Showing wrinkles disappearing by using a cream through before-and-after comparisons". The format itself survives: "Comparative claims may be allowed, subjected to the provision of evidence or clear disclaimer(s)." What is disallowed is "Product effect comparisons, such as before-and-after results, which may cause viewers to have a false or distorted impression about a product's outcome/results".

Healthcare and Pharmaceuticals(August 2026) is a table of 45 market sections, not one rule. The cosmetics rule most blogs quote as global is "Do not make medical claims", with the examples "Exaggerated performance promises concerning a product's effect" and "Claims to remove or cure acne, remove wrinkles, or treat hair loss". It appears in 17 of the 45 sections, including the United Kingdom, France, Italy, Spain and Israel. The United States section has no cosmetics row and no supplements row. "Before-and-after comparisons for over-the-counter medicines, medical devices, and food supplements" is listed as not allowed in six sections: the Gulf and North Africa group (Bahrain, Egypt, Iraq, Kuwait, Morocco, Oman, Saudi Arabia), Jordan, Lebanon, Pakistan, Turkey, and the Qatar, South Africa and United Arab Emirates section. One creative run worldwide is governed by its strictest row.

Weight Management(September 2026) is global: "Any weight loss or muscle gain claims must be restricted to audiences aged 18 years or older, and must do so via the promotion of a healthy lifestyle." Not allowed: "Claims that the product alone, without diet or exercise, can lead to weight loss or muscle gain." and "Suggestions that losing weight or gaining muscles is easy or guaranteed." Weight loss supplements are listed for the United States and Canada as "what may be allowed subject to age restrictions, applicable local licensing requirements, and full compliance with all advertising policies".

Rulebook three: TikTok Shop wrote rules for the AI actor itself

When the same video is attached to a product in TikTok Shop, the TikTok Shop Academy policy pages apply, and they are the only ones written with a synthetic presenter in mind. The US AIGC policy (09/02/2026) and the UK AIGC policy (28/08/2026) share a structure and differ in detail. The whole of that policy, including the product-in-frame and LIVE rules, is in TikTok Shop's AI content rules; this section takes only the health claims.

The presenter's costume is a claim.The US page lists as a violation "Impersonating or fabricating professional images such as 'doctors' or 'professors' to promote health-related products (including food, dietary supplements, medical devices, health services, etc.)" and adds: "Such behavior, especially in health-related fields like dietary supplements and functional foods, will be a key focus of governance." The UK page gives the worked example, "A Short Video with an AI Doctor dressed in scrubs promoting a supplement saying 'I take one of these capsules every morning'", and states: "This is a serious violation of our policy and may result in removal of e-commerce permissions." The banned personas run from dermatologists to "fitness coaches, beauty specialists, wellness experts".

The visual is a claim.UK: "AI must not be used to demonstrate or imply that a product can cure, heal, eliminate, treat, prevent, eradicate, or alleviate any disease or medical condition, or health condition, even if no explicit text claim is made." US: "Before-and-after visuals must not compress time or imply instant results unless the product can support that claim with reliable evidence." and "AIGC must not be used to demonstrate a core product effect that is not supported by the actual product, product detail page, instructions, ratings, certification, test results, or other reliable evidence." That last sentence is the FTC's substantiation standard rewritten for a video generator.

The gut-cleanse animation is banned by name.US: "Do not use AI-generated scary or shocking images (such as damaged organs or diseases) to unduly influence people to buy supplements or health products or to mislead them about the product's effectiveness." UK, as a non-compliant example: "Shock imagery showing a gut full of 'waste', toxins, or parasites, followed by taking a supplement and an unrealistic or instantaneous expulsion of bodily waste".

The penalties differ too. The US tier for repeat offenders "may include limiting posting capabilities, permanently disabling commission withdrawal privileges, or suspending or banning the account entirely." Ads Manager rejects an ad; Shop can freeze the money.

Six script lines through all three rulebooks

The FTC column is about evidence, the platform column about wording and targeting, the Shop column about what the AI shows. A script can pass the first two and fail the third on a visual added to make the hook land.

What changes when the presenter is generated

Nothing in rulebooks one and two. The FTC does not care whether the person saying "it worked for me" exists; it cares whether the product works for the typical buyer. Meta and TikTok Ads apply their claim rules to words and imagery, and neither policy quoted above softens or tightens a claim rule because the presenter is generated. TikTok's Misleading and False Content policy does add a disclosure requirement for AI footage, and that separate question is covered in Meta's AI creative policy and TikTok's AI disclosure rules.

What changes is how easy the violation is to produce. A human creator hesitates before putting on scrubs or animating a colon. A video model renders "dermatologist in a white coat holding the serum" without a flicker, and the talking-head models we run (OmniHuman 1.5 and VEED Fabric 1.0, driven by a voice track generated from your script, and Pruna Avatar, which speaks the script with its own built-in voice) speak whatever script they are handed. Our pipeline passes your prompt and script to the model unchanged; there is no claim filter in the render step, and you should assume no tool has one. On Pruna Avatar, which we bill at 18.4 credits per spoken second, a 30-second supplement testimonial is roughly 550 credits. If the ad is bounced for a line you could have cut, those credits bought a rejection notice. So make the three edits on the script, then render: qualify or delete every timeframe, strip scrubs and titles from the presenter, and keep the before-and-after but keep its clock honest. The skincare and supplement scripts in our UGC ad script templates are written that way: routine and texture, not outcomes.

None of this is legal advice, and every page above changes without notice. Meta's change log carried an entry dated the day we read it, which is the best argument for checking the primary page before trusting any summary, including this one.

Sources

• Federal Trade Commission, Health Products Compliance Guidance (December 2022), sections II.C, III.B and III.C.1, Example 41
• Meta Transparency Center, Advertising Standards, Health and Wellness (read in-browser 4 September 2026; change log entries dated that day, 23 July 2026 and 27 December 2024)
• TikTok Advertising Policies, Misleading and False Content (last updated April 2026)
• TikTok Advertising Policies, Healthcare and Pharmaceuticals (last updated August 2026), 45 market sections counted on that page
• TikTok Advertising Policies, Weight Management (last updated September 2026)
• TikTok for Business, Health Advertising Guide (marketing guidance, cited only for its disclaimer suggestions)
• TikTok Shop Academy (US), AI-Generated Content Restrictions and Requirements (09/02/2026)
• TikTok Shop Academy (UK), Artificial Intelligence Generated Content (AIGC) (28/08/2026)

All eight were read directly on 4 September 2026. The dates in brackets are how you check whether this post has gone stale.

The claim rules are one half of the pre-purchase question; the other half is whether the platform labels or rejects the AI actor itself. Meta labels detected AI on its own, while TikTok rejects undisclosed AI footage. If the script is clean and the question is which model speaks it best, the lip-sync comparison covers the audio-driven models. Once the script passes all three, UGC Vids AI is where it gets rendered: plans start at $49 a month and the first 3 days are free.