Quick answer: Yes, a brand can run AI UGC and AI influencers in the US, UK and EU, but not as fake customers. New York has required a conspicuous disclosure of synthetic performers in ads since 9 June 2026. The FTC has no blanket ban on AI avatars but prohibits fake or false testimonials. The EU has required visible deepfake labels since 2 August 2026, and the UK has no blanket label rule but bans ads that mislead.
Rules below were checked at the source on 1 November 2026. One bias up front: we run a marketplace that books human creators, so treat our view on AI versus humans as an interested party's. Our explainer on what a UGC creator is covers the wider AI versus human debate, so this piece stays on disclosure.
AI UGC disclosure rules by jurisdiction
| Where | What it covers | What you must do | Consequence |
|---|---|---|---|
| New York, from 9 June 2026 | Visual or audiovisual ads with a synthetic performer, where the advertiser has actual knowledge | Conspicuously disclose it in the ad. The law sets no wording or size | $1,000 first violation, $5,000 after that |
| USA, FTC | Fake or false reviews and testimonials, plus deceptive ads generally | Never misrepresent that the testimonialist exists, used the product, or had that experience | Up to $53,088 per violation, for knowing violations |
| UK, ASA and CMA | All ads under the CAP Code. Fake reviews under consumer law | No blanket AI label. Disclose where the audience would otherwise be misled, and an AI influencer is the named example | ASA rulings. The CMA can fine up to 10% of global turnover |
| EU, from 2 August 2026 | Deployers of AI that generates deepfakes, including outside the EU where the output is used in the EU | Clear, visible label at first exposure at the latest | Up to 15 million euros or 3% of worldwide turnover |
| Meta (platform rule) | Ads on its apps | Self-disclosure for social issue, election and political ads. We found no such duty for other ads | Meta applies an AI info label, with uneven coverage |
We could not open TikTok's ad policy pages on 1 November 2026, so we make no claim about them. Check inside your ad account before you run there.
New York's synthetic performer law
We read the bill itself, S8420-A, signed on 11 December 2025. It amends General Business Law section 396-b and takes effect on the 180th day after becoming law, which lands on 9 June 2026. Here is what it says.
- Who: anyone dealing in property or services who, for a commercial purpose, produces or creates an ad, in any medium.
- What: a synthetic performer is a digitally created asset, made using generative AI or a software algorithm, meant to give the impression of a human performer who is not recognizable as any identifiable real performer. So an invented person is covered. A clone of a real person is outside this definition, and the bill keeps other rights such as Civil Rights Law section 50-f intact.
- Knowledge: the duty applies where the person has actual knowledge. If you generated the avatar, you know.
- Not covered: audio ads, ads where AI only translates a human performer's language, ads for expressive works such as films and games when the use matches the work, and the media that merely carry the ad.
Two gaps matter in practice. First, the law does not define conspicuous. Kelley Drye, a law firm, wrote on 8 June 2026 that the law doesn't elaborate and no guidance exists, and we found none as of 1 November 2026. The same note says there is no private right of action. Second, subdivision 2 of that section is limited to ads in this state, and the new subdivision 3 has no such words. We found nothing settling how far the duty reaches, so if your ad serves in New York, treat it as covered. Check your setup with a lawyer.
FTC rules on fake reviews and AI testimonials
The FTC's final Consumer Review Rule says "AI-generated reviews are covered by the final rule". It also limited the ban to material misrepresentations, so that non-deceptive speech such as virtual influencers, computer-generated fictional characters, is not caught. The FTC's published questions and answers on the rule, dated November 2024, address AI stock avatars by name. There is no blanket ban, and stock avatars are not giving consumer reviews. But an avatar might count as a testimonial, which is prohibited if the underlying testimonial is fake or false. The FTC adds that avatars could also be deceptive under the FTC Act, and that a celebrity avatar used without permission would violate the rule.
The rule text bars a business from spreading a testimonial it knew or should have known misrepresents that the person exists, that they used the product, or what their experience was. The Endorsement Guides go further: when an ad says the endorser uses the product, the endorser "must have been a bona fide user of it". An AI presenter has never used anything. So an avatar can present your product, but it cannot claim a month of personal use.
The FTC's penalty table sets the maximum for rule violations at $53,088, and the final rule says courts weigh statutory factors and may impose much less per violation. Civil penalties depend on knowledge. Our guide to FTC rules for UGC creators covers paid-partnership disclosure, which still applies on top. Ask a lawyer before running any AI testimonial.
Need someone who can honestly say they used it? Post your brief and let creators apply. Creators show verified stats, and payment is held until the work is approved.
Post your briefEU rules from August 2026
The Commission's FAQ on Article 50 of the AI Act says the transparency rules apply from 2 August 2026. A grace period to 2 December 2026 covers only the machine-readable marking duty on providers of older systems, not you. Content generated before 2 August 2026 does not need labelling retroactively.
Your duty is as a deployer, which includes a company using AI in its business. A company stays the deployer even if contractors run the system on its behalf and under its control. Deployers must label deepfakes, and the FAQ sets three cumulative tests: a high level of similarity to the subject, a subject that exists or can plausibly exist, and a false appearance of being authentic. Our reading is that a photorealistic AI customer built to look like real UGC likely meets all three, and a cartoon mascot does not.
The label must be clear and visible at first exposure at the latest. A hidden machine-readable mark does not satisfy it. Article 2 reaches deployers outside the EU where the output is used in the EU. So a US brand running this creative to EU audiences is in scope. Have a lawyer in your target market confirm your reading.
UK guidance on AI ads and AI influencers
The UK has no AI-specific ad rule. The Advertising Standards Authority said in May 2025: "There is no blanket legal requirement in the UK to disclose the use of AI in ads". Its test is whether the audience is likely to be misled without disclosure. It says making clear "that an influencer is AI-generated, could well help to negate an otherwise misleading impression". It also says a label will not rescue a false claim, using a cosmetic ad whose AI after image does not reflect real results as its example.
In June 2026 it added that an AI celebrity endorsement realistic enough to mislead is likely to breach the Code. The October 2026 influencers' guide says AI content must also comply, and that platforms may require disclosure.
Fake reviews are a separate, harder line. The Competition and Markets Authority's guidance defines a fake review as one that purports to be, but is not, based on a person's genuine experience, and lists software-generated reviews as a banned example. The CMA can fine up to 10% of global turnover. A UK lawyer can tell you which side of that line your ad falls on.
Meta's AI label is not your disclosure
Meta's newsroom post, updated 1 June 2026, says it labels ads significantly edited with its own generative tools, behind the three-dot menu or next to the Sponsored label, and always next to Sponsored when those tools add an AI photorealistic human. It also says it is beginning to detect third-party AI tools automatically and show the label under About this ad. Yet Meta's help page on AI images in ads, as we read it on 1 November 2026, still listed third-party tools as ones that won't get the label.
So Meta's own pages disagree, and either way the label depends on detection and may sit behind a menu. Treat it as extra. Put your own disclosure in the video.
Disclosure wording template
This is our suggested wording, not a safe harbor. No regulator above has approved a set phrase.
| Your ad | On-screen text | First line of the caption |
|---|---|---|
| AI presenter, scripted product demo | AI-generated presenter | Ad. The person in this video is AI-generated, not a real customer. |
| AI presenter reading a real customer's words | AI presenter. Real review, used with permission | Ad. AI-generated presenter reading a review from [first name, month, year]. |
| Brand-run AI influencer account | Virtual influencer. Not a real person | Ad for [brand]. [Name] is an AI-generated character. |
| Real creator, AI-made scenes | Includes AI-generated scenes | Ad. Filmed by [creator]. Some scenes are AI-generated. |
Four rules for using it:
- Put it in the video, not only the caption. The FTC's Endorsement Guides define clear and conspicuous as difficult to miss, and judge a visual disclosure by size, contrast, location and time on screen. That is the FTC's definition, not New York's, but it is the best benchmark we found.
- Show it from the first frame and for as long as the AI person is on screen.
- Remove first-person experience lines from the script, such as I use this or my skin changed, and any AI before and after image.
- Keep a file: the tool used, who generated it, the source of each claim, and permission for any real review. Add disclosure to your creative brief so vendors deliver it that way.
Where AI falls short of real creators
- It cannot give a lived testimonial. That is the bona fide user rule above. In skincare, health or finance, the personal claim is often the ad.
- You may not own it. The US Copyright Office's January 2025 report concludes copyright does not extend to purely AI-generated material, and that prompts alone do not give enough control. That is its position, not a court ruling. Compare that with usage rights on creator content, which you can license in writing.
- Consumers like it less than ad teams think. IAB's January 2026 survey of more than 500 Gen Z and Millennial consumers and 100 ad executives found 82% of executives believe those consumers feel positive about AI ads. Only 45% do.
- Disclosure may cost clicks. IAB's respondents said awareness could raise purchase likelihood. Yet PPC Land, a trade publication, reported NYU Stern research where an AI disclosure cut click-through rate by 31.5%. We did not read the study itself. Test it on your own account, but do not drop a required label because it hurts.
We found no independent study comparing sales from AI and human UGC, so we won't claim one wins. A sensible split: labelled AI for product demos and hook tests, real people for anything that depends on experience. Our guide to hiring UGC creators covers the human side.
Splitting budget between AI tests and real creators? Map the spend before you commit to either.
Open the campaign budget plannerFAQ
Do I have to disclose AI UGC ads in the US?
In New York, yes, if a synthetic performer appears and you have actual knowledge. We did not review other states. Federally, the FTC's concern is deception: fake testimonials are prohibited whether or not you label them.
Are AI influencers legal?
Yes, with limits. The FTC says its rule was drafted not to prohibit virtual influencers, and the ASA treats disclosure as a way to avoid misleading people. The false claims rules still apply to whatever the character says.
Does New York's law apply if my brand is not in New York?
We found no guidance settling this. The new subdivision does not limit itself to ads in this state. If your ad serves there, treat it as covered and ask a lawyer.
Sources
- New York State Senate: Bill S8420-A
- Kelley Drye (law firm commentary): NY synthetic performer law goes into effect
- Federal Register: FTC final rule on consumer reviews and testimonials
- Cornell Legal Information Institute: 16 CFR 1.98
- Cornell Legal Information Institute: 16 CFR 465.2
- Cornell Legal Information Institute: 16 CFR 255.1
- Cornell Legal Information Institute: 16 CFR 255.0
- European Commission: Transparency obligations under Article 50 of the AI Act
- EU AI Act Explorer: Article 2
- ASA: Disclosure of AI in advertising
- ASA: AI and deepfakes, four things advertisers need to know
- ASA and CMA: Influencers' guide, October 2026
- CMA: Fake reviews guidance (CMA208)
- GOV.UK: New consumer protection regime comes into force
- Meta: Expanding GenAI transparency for Meta's ads products
- US Copyright Office: Copyright and AI, Part 2, Copyrightability
- IAB: The AI Ad Gap Widens
- PPC Land (trade press): AI ad labels cut clicks 31.5%
