Labelled AI, every time.
Every creator on Synthicate is an AI creator and is labelled as one, on every card, page, offer and sponsored post. These rules are in the product, not in a policy PDF.
The short version
- Every creator is labelled AI, on every card, page and offer, and its studio confirms its public profiles say so.
- Every sponsored post carries three labels: the platform’s paid-partnership label or #ad, its AI-generated label, and “AI-generated” on screen in the video itself.
- No first-hand testimonials, and some categories can’t be advertised at all.
- Synthicate never posts, schedules or automates anything on a studio’s accounts.
Every creator is labelled AI
The label is on every creator’s card, page and offer in the app. It isn’t a setting, and it can’t be switched off.
Before brands can see a creator, its studio confirms two things: that the creator’s public profiles say it’s AI, and that it isn’t modelled on a real, identifiable person, with the studio owning the rights to its look and voice. On Instagram, the account also carries Instagram’s “AI-generated profile” label.
Three labels on every sponsored post
Every sponsored post carries the platform’s paid-partnership label, or #ad up front, the platform’s AI-generated label, and “AI-generated” on screen in the video itself, so the label goes wherever the clip is shared. The studio ticks all three when it sends the live link, and the brand sees the ticks.
#ad or the platform's paid-partnership label, up front
The platform's AI-generated label switched on
"AI-generated" on screen in the video, for its first 3 seconds
All three are written into every product pack, beside the testimonial rule, and can’t be removed.
No first-hand testimonials
A creator that doesn’t exist can’t have used anything, so it can’t say it did. First-hand reviews, results and before-and-afters are out.
The brief checker reads a brief as the brand types it, flags lines like these and suggests a rewrite. A first-hand line blocks: the brief can’t be saved until it’s rewritten. A health claim or a guarantee only warns, because the brand may hold the evidence, and the studio sees the warning on the brief.
Flagged: She’s been using it for weeks
Rewritten: She shows how the serum fits into a night routine.
Blocked categories
Some products can’t be advertised on Synthicate at all, and a product pack can’t be made for them. Each has its own Australian rulebook; some may reopen once the product can enforce it.
- Therapeutic goods: supplements, sunscreen, medicines
- Australia bans paid testimonials for therapeutic goods outright.
- Alcohol
- Needs 25+ looks, a 75%+ adult audience and age-gated posts (ABAC).
- Gambling and betting
- Separate state rules and platform bans.
- Tobacco, vapes, nicotine
- Advertising is banned in Australia.
- Financial products and crypto
- Finfluencer licensing rules; YouTube demonetises AI money advice.
- Weapons, adult, political
- Not on Synthicate.
Categories that need care
Other products can be advertised, with care. A product pack in one of these categories starts with its rules already written in.
Skincare and cosmetics
A claim to treat a condition makes it a therapeutic good, which is blocked.
Must: product in frame for 3 s or more, label to camera.
Never: "Clinically proven" unless the brand supplies the evidence; change the pack colours or the logo.
Fashion and accessories
Fit and colour must match the photos.
Must: colours true to the reference photos.
Never: invent sizes, fabrics or prices.
Food and drink (no alcohol)
No health claims beyond the label.
Must: show the pack as sold.
Never: health claims not on the label; aim the post at children.
Tech and gadgets
Screens show only real features.
Must: real screens and real features only.
Never: invent specs, battery life or benchmark numbers.
Home and living
True scale, true colours.
Must: true scale next to everyday objects.
Never: show it bigger or smaller than it is.
Apps and games
Real footage only.
Must: real app or game footage.
Never: fake gameplay, fake payouts or fake reviews.
Travel and experiences
The creator hasn't been there.
Must: real photos of the place or service.
Never: claim the creator went there.
Pets
No claims for pet medicines.
Must: show the product as sold.
Never: health claims for pets.
Fitness gear (not supplements)
Gear only; supplements are therapeutic goods.
Must: show the gear in use.
Never: body-transformation claims.
Never on Synthicate
- A creator that pretends to be a real person
- A creator modelled on a real, identifiable person
- First-hand reviews, results or before-and-afters
- Sponsored posts without the AI and paid labels
- Selling fake followers, views or engagement
- Deals paid outside the platform after meeting on it
- Adult content, in any category
- Ads aimed at children
We never post on your accounts
Synthicate never posts, schedules or automates anything on a studio’s accounts. The studio posts on its own account, with all three labels, and pastes the live link into the deal.
TikTok treats automation across many accounts as spam, and can ban every account an operator runs. Staying off a studio’s accounts keeps its roster safe.
What the rules say
What the rules above rest on, from regulators and the platforms. Each row links its sources.
FTC
Labelled virtual influencers are legal: the reviews rule was drafted so as not to prohibit them. An endorser must have really used the product, so an AI creator can’t say it has.
ACCC
Advertising must be clearly distinguishable as advertising. The ACCC fined its first business over influencer disclosure in March 2026.
TGA
No paid testimonials for therapeutic goods, disclosed or not.
ABAC
Alcohol ads can’t use anyone who is or looks under 25, and need an audience that’s 75% or more adult.
TikTok
Realistic AI content must be labelled, and using automation to run many accounts is spam.
Accounts featuring AI-generated people need the “AI-generated profile” label, and sponsored posts still need the Paid Partnership tool.
YouTube
Realistic synthetic content must be disclosed, and AI creators talking health, money or law are demonetised.
EU AI Act, Article 50
From 2 August 2026, deepfakes must be disclosed “upon first exposure at the latest”.
Read on 29 Sep 2026. Every source the rules rest on, grouped: