Academy

/

AI-Generated "Synthetic UGC" Ads: When a Brand Uses an AI Actor Instead of a Real Creator

Guide

Strategy

Brands

AI-Generated "Synthetic UGC" Ads: When a Brand Uses an AI Actor Instead of a Real Creator

A synthetic UGC ad is a video advert that looks like organic user-generated content but is entirely AI-generated: a tool like Arcads or HeyGen turns a script into a photorealistic AI actor delivering it, with no real creator, filming or customer involved. It's often cheaper and faster than hiring a real UGC creator, but it triggers three separate disclosure duties (Meta, TikTok and the EU AI Act) and can cross into deceptive-testimonial territory under the US FTC's rule if it's styled as a genuine personal experience.

Short answer: a real-looking ad with no real person behind it

A synthetic UGC ad is a video advert built to look like organic, user-generated content — handheld camera, direct-to-camera delivery, testimonial style — except the entire thing is AI-generated. Tools like Arcads and HeyGen take a script and have a photorealistic AI actor from a library deliver it, with an AI voice and lip-sync. There's no real creator, no real camera and no customer who has actually used the product. The format has grown fast through 2025-2026 because it solves a concrete production problem — a shortage of ad creative — markedly cheaper and faster than hiring and briefing real people.

What a synthetic UGC ad actually is

The mechanics are the same across tools: you (or the tool's built-in script generator) write a script, pick an AI actor from a library — Arcads' own site advertises over 1,000 AI actors to choose from — and the tool renders a video of that actor speaking the script with a generated voice and synced lip movement. Some tools, HeyGen among them, also offer avatar cloning, letting a brand build its own recurring AI figure from real footage it supplies.

It's worth keeping the format distinct from three things it's often confused with:

  • Real UGC. See what is UGC and UGC creator vs influencer — both cover content made by a real person who actually tried the product, whether or not that person posts it themselves or just delivers the file.
  • A virtual/AI influencer. An AI actor in a synthetic UGC ad is single-use for one campaign, with no name, no account of its own and no continuing "personality" across posts. A virtual influencer like Lil Miquela or Aitana López is the opposite: a fixed, named, recurring persona with its own profile. See AI and virtual influencers vs human creators for the full category — and for the same disclosure rules this article applies to a one-off ad format instead.
  • Meta AI Studio's Creator AI. That's a real creator's own AI chatbot extension of themselves, trained on their own content, always labelled as AI. See Meta AI Studio: Creator AI personas explained. A synthetic UGC ad has no real person behind it at all.

Synthetic UGC ad vs the three alternatives

AspectSynthetic UGC adReal UGC (hired creator)Virtual/AI influencer
Who's talking?An unnamed AI actor, used for one adA real person who tried the productA fixed, named, recurring AI persona
Genuine product experience?No — the script can claim an experience nobody hadYes, in principleNo
Production timeHoursDays to weeksWeeks to months (set-up) + ongoing production
Scales across languages/variants?Very easily — new script, same actorNeeds a new creator or shoot per marketUnlimited in theory, but needs new production per post
Disclosure dutyMeta, TikTok and the EU AI Act — see belowOrdinary ad-disclosure rules (a human is behind it)Same three duties as synthetic UGC, plus ongoing labelling upkeep across every post
Risk of reading as a fake testimonialHigh, if the script is written in first person as a personal experienceLow, if the experience is genuineLow — doesn't typically present as a "customer"

The rules that actually apply

Meta: the AI Info label — a stricter requirement for photorealistic AI humans

Meta states that ads "created or significantly edited" with generative AI tools get an "AI Info" label in "About this ad." Where the label appears depends on the content: if the ad includes a photorealistic AI human — which a synthetic UGC ad is by definition — Meta places the label directly next to "Sponsored," not tucked into the three-dot menu, which is the rule for less visible AI edits. Meta has also said it automatically detects content made with third-party AI tools via industry-standard signals (typically C2PA metadata embedded in the file) and applies the label itself — meaning an ad can end up labelled even if the advertiser never self-disclosed it.

TikTok: labelling "realistic-appearing" AI content

TikTok's own synthetic-media policy requires labelling AI-generated content that contains "realistic images, audio or video" — exactly the category a photorealistic AI actor falls into. TikTok offers both a self-serve labelling tool and is testing automatic detection of unlabelled AI content; content that goes unlabelled and is judged misleading can be removed under TikTok's own rules. TikTok's own Symphony suite, by contrast, automatically labels its own AI-generated output — see TikTok Symphony: TikTok's own AI ad creation suite and digital avatars for how that compares to third-party tools like Arcads and HeyGen.

The EU AI Act: applies to a one-off ad too, not just a persistent persona

Article 50 of the EU AI Act distinguishes two duties already covered in the virtual influencers article — and they apply just as much to a single synthetic UGC ad as to a persistent persona. Article 50(2) requires the provider of the AI system (in practice, tools like Arcads/HeyGen, or the brand if it builds its own solution) to mark the output in a machine-readable format as artificially generated — whether the actor appears once or a hundred times. Article 50(4) requires the deployer — the brand publishing the content — to disclose to the audience if the content qualifies as a "deepfake": synthetic image or video content that resembles an existing or plausibly existing person. A photorealistic AI actor that looks and talks like an ordinary human meets that definition just as precisely as a persistent virtual influencer does — even though the actor never appears again.

Ordinary ad-disclosure law still applies, AI or not

The AI Act's labelling requirement doesn't replace the ordinary Danish and EU rule that commercial content must be marked as advertising under Markedsføringsloven § 6(4). See the full breakdown in influencer marketing disclosure rules in Denmark and the EU. A synthetic UGC ad has to satisfy both rule sets at once, not one instead of the other.

The risk most brands overlook: fake testimonials (the FTC, US)

The US Federal Trade Commission's Consumer Reviews and Testimonials Rule took effect on 21 October 2024. It bans businesses from creating or promoting fake consumer or celebrity testimonials that misrepresent the identity, experience or existence of the reviewer — and the FTC's own material explicitly names AI-generated reviews and testimonials as an example of what the rule covers. Legal analysis of the rule (including from the law firm Sidley Austin) assesses that an AI-generated avatar delivering testimonial-style advertising would likely fall within that prohibition — the rule doesn't explicitly address this exact scenario, but the logic is the same: a claimed personal experience that never happened. Violations can carry civil penalties under the rule — some legal trackers report a figure around $51,744 per violation based on the rule's level at its 2024 introduction, but FTC penalty amounts adjust periodically for inflation, so the exact figure should be checked before it's used in an actual advisory conversation.

The practical point: the risk rises sharply when the script is written in first person as a genuine personal experience ("I've been using this for three weeks and my skin has never looked better"). A synthetic UGC ad that demonstrates the product instead, without claiming a personal experience, sits outside the rule's core target. There's no direct Danish or EU equivalent to the FTC's specific rule, but the UCPD's general prohibition on misleading commercial practices points the same direction: a falsely claimed personal experience is misleading information about the product, whether the applicable law is American or Danish. That's a logical parallel, not a documented Danish enforcement case.

Do the tools handle disclosure for you?

Several secondary sources suggest tools like Arcads and HeyGen "bake in" a disclosure label or watermark automatically. We couldn't confirm that: a direct check of Arcads' own website and HeyGen's own UGC ad-maker page found no mention whatsoever of built-in AI labelling, watermarks, or compliance with Meta's, TikTok's or the FTC's rules. Arcads also has no public pricing page — the old arcads.ai/pricing URL returns a 404, and pricing only appears once you create an account. The conclusion: responsibility for labelling and compliance sits with the brand commissioning and publishing the ad — not with the tool, unless a vendor's own site explicitly documents otherwise at the time you use it.

Decision framework

IF the ad demonstrates the product without claiming a personal experience → lower FTC risk, but Meta/TikTok labelling and the AI Act's Article 50 still apply.

IF the script is written as a first-person personal testimonial ("I've used...") → the highest risk under the US FTC rule, and the least forgivable mistake to be caught in.

IF you need many language or hook variants of the same message fast → that's the format's genuine strength, and where it typically makes the most sense.

IF you want a fixed, recognisable "face" across many campaigns over time → that belongs to the virtual-influencer category, not synthetic UGC — see AI and virtual influencers vs human creators.

IF budget allows it and you value the — limited but real — trust a real person adds → a real UGC creator or influencer still wins on authenticity, see UGC creator vs influencer.

Worked example (hypothetical, illustration only)

The figures below are entirely hypothetical and only illustrate the price-versus-speed trade-off — this is not a claim about actual market prices, and not a real Make Influence customer case.

A brand needs 10 script variants for a paid social test. With real UGC creators at an illustrative $140 per delivered video (see how much does UGC cost for actual market pricing), that's 10 × $140 = $1,400, with a turnaround of several days to weeks. With a synthetic-UGC tool at an illustrative $14 per generated video, that's 10 × $14 = $140, with turnaround measured in hours. The gap illustrates why brands test this route — but the maths doesn't include the time it takes to make sure all 10 ads are correctly labelled on every platform they run on, which doesn't scale the same way as the production itself.

Common mistakes

  • Assuming the tool handles labelling for you. Neither Arcads' nor HeyGen's own site promises that — always check a vendor's own documentation before relying on a claim of automatic compliance.
  • Writing the script as a genuine personal experience without marking it as AI. That's exactly the combination the US FTC's rule is built to catch.
  • Assuming the AI Act only applies to persistent personas. Article 50 applies just as much to an AI actor that appears in a single ad.
  • Overlooking that Meta and TikTok can label the content automatically even if you didn't. Plan for the label to appear regardless of whether you set it yourself.
  • Comparing the price without counting the compliance work. The low per-video price doesn't include automatic labelling and approval across every platform.

Make Influence's operational perspective

Make Influence's model is built for collaborations with real, external creators — not for building or recommending AI actors. When brands ask us about synthetic UGC, our advice is to treat it as another production tool, not a substitute for a UGC creator: it solves a volume-and-speed problem in ad creative, but it doesn't solve the trust a real person with a real experience adds. The labelling and claims discipline needs to be at least as tight as if it were a real human — arguably tighter, because three separate rule sets are watching. This is our operational view, not a claim that the format never makes sense.

FAQ

Is a synthetic UGC ad the same as a virtual influencer?

No. A synthetic UGC ad uses an unnamed AI actor for a single ad with no continuing persona. A virtual influencer like Lil Miquela is a fixed, named figure with its own profile across many posts. See AI and virtual influencers vs human creators.

Do you always have to label a synthetic UGC ad as AI-generated?

On Meta and TikTok, yes, if it runs as a paid ad — both platforms require it, and can detect it automatically in practice. In the EU, a photorealistic AI actor likely also triggers the AI Act's Article 50(4), because it resembles a plausibly existing person.

Do Arcads or HeyGen handle disclosure for you?

Not according to their own websites at the time we checked. Neither Arcads' nor HeyGen's UGC ad page mentions built-in AI labelling or platform compliance as a feature. Assume the responsibility sits with you.

Can a synthetic UGC ad be illegal in the US?

The ad format itself isn't illegal. The risk arises if the script is written as a genuine personal customer testimonial without disclosing that the speaker is an AI actor with no real experience — that's exactly what the FTC's Consumer Reviews and Testimonials Rule, in effect since 21 October 2024, prohibits.

Is it cheaper than hiring a real UGC creator?

According to secondary sources, yes, often markedly — but Arcads publishes no pricing list of its own, so the specific figures aren't vendor-confirmed. See how much does UGC cost for actual market pricing on real UGC.

Does Make Influence use AI actors in campaigns?

No. Make Influence's model is built for collaborations with real, external creators. See Make Influence's operational perspective above.

Is TikTok's own Symphony tool the same as Arcads or HeyGen?

No. Symphony is TikTok's own, free, built-in AI toolkit inside TikTok Ads Manager, and its Creative Studio output is automatically labeled "AI-generated" — a labeling behavior neither Arcads nor HeyGen documents on their own sites. See TikTok Symphony: TikTok's own AI ad creation suite and digital avatars for the full comparison.

Does Content Credentials (C2PA) affect a synthetic UGC ad?

It can. Some third-party AI actor tools may embed C2PA metadata in their output, and platforms that read it (TikTok, Meta) can label the ad automatically as a result — on top of, not instead of, the platform's own labelling rules described above. See Content Credentials (C2PA) explained for how that technical standard works and its limits.

Make Influence

Want influencer marketing to be easier?

Find creators with real audience data, run collaborations in one place, and see clicks and sales per creator while the campaign is live.

Book a demoCreate account

Make Influence

Get paid for the audience you built

Apply to campaigns from brands that are actively looking, follow your own clicks and sales, and get paid without chasing invoices.

Create creator profileMore creator guides

Make Influence

One place for the whole collaboration

Briefs, agreed terms, tracking links and results sit together — so brands and creators see the same numbers.

See how it worksBrowse the Academy