AI-generated UGC — the shaky-handheld, first-person, "just a real customer talking to their phone" ad, made with a synthetic actor and a script — became the dominant performance-creative format of 2026 because it is cheap to produce, fast to iterate, and it converts. Its defining trait is also its regulatory problem: it works precisely because it looks like an unpaid, unscripted, genuine person, when it is none of those things. That collision is what Meta's clearer AI ad labeling is a response to. Meta applies an "AI info" label to ad creative it detects as AI-generated or that was made with its own generative tools, requires advertisers of social, electoral, and political ads to self-disclose AI-created or -altered photorealistic content, and from June 1, 2026 runs automated detection that labels third-party AI media in ads with no advertiser action. Layered on top is the FTC's 2024 rule banning fake and AI-generated testimonials outright — the exact deceptive-endorsement risk an AI UGC ad can trip if the synthetic person is presented as a real, satisfied customer. This guide separates the two things people conflate: disclosure (telling viewers the media was made with AI, which is a labeling task) and deception (passing off a fabricated person as a genuine one, which is a legal line). It covers how Meta's labels actually work, what the FTC rule prohibits, why the UGC-style format is the sharp end of both, and how to run AI UGC as a durable format instead of a liability.
The most effective performance-ad format of 2026 is also the one regulators are circling: AI-generated UGC. It is the vertical, handheld, first-person clip that looks like a real customer talking to their phone about a product — except the customer is a synthetic actor and the script is generated. It wins because it is cheap to produce, trivial to iterate into fifty variants, and it converts better than polished brand film precisely because it reads as unproduced and genuine. That last property is the whole problem. The format works by looking like an unpaid, unscripted, real person, and it is none of those things. The rise of the format itself is covered in [AI UGC ads: the rise of synthetic creator-style ads as a performance format](/guides/ai-ugc-ads); this guide is about the disclosure and honesty layer that decides whether the format survives contact with the rules.
Two separate systems govern it, and conflating them is the most common and most expensive mistake. The first is disclosure: telling the viewer the media was made with AI. That is a labeling task, and it is what Meta's clearer AI ad labels — the "AI info" marker, the political-ad self-certification, and the automated detection that expands on June 1, 2026 — are about. The second is deception: presenting a fabricated person as a real, genuine customer. That is a legal line drawn by the FTC's truth-in-advertising authority and its 2024 rule banning fake and AI-generated testimonials. An ad can be fully labeled as AI-made and still be an illegal fake testimonial. Labeling the pixels does not fix a counterfeit endorsement. This guide covers how Meta's labels actually work, what the FTC rule prohibits, why UGC-style creative is the sharp end of both, and how to run the format as a durable channel instead of a liability.
Meta's AI labeling for advertising has three moving parts, and it helps to see them as distinct rather than one blanket rule. The first is the "AI info" label itself. When Meta detects AI-generated or AI-edited media in an ad, or the ad was built with Meta's own generative creative features, it attaches an "AI info" marker that shows on the "About this ad" screen reachable from the three-dot menu, and sometimes next to the Sponsored label at the top of the ad. The trigger is meaningful edits — generating the visual subject, background generation, compositing a scene — not cosmetic ones. Cropping, resizing, and color correction do not set it off. An advertiser cannot remove a label Meta has applied automatically.
The second part is self-certification, and it applies to a specific, higher-stakes slice: ads about social issues, elections, or politics. For those, Meta already requires the advertiser to proactively disclose when the ad uses AI to create or alter photorealistic image, video, or audio that depicts a real person saying or doing something they did not, that fabricates a realistic-looking person or event, or that presents an unverified event as a real recording. This is an affirmative obligation on the advertiser, not something Meta infers. Failing to disclose when required gets the ad rejected, and repeated failure can bring penalties. Ordinary commercial ads do not carry this same universal manual toggle, which is exactly why the boundary between "stylized product ad" and "ad about a social issue" matters.
The third part is automated detection, and its scope widened on a firm date: starting June 1, 2026, Meta uses automated technology to independently identify ad media created or edited with third-party generative AI tools, and applies the "AI info" label to it with no action needed from the advertiser. This is the piece that closes the gap where an advertiser used an outside tool and said nothing. Detection leans on provenance signals — the C2PA and IPTC metadata that many generators now embed, part of the same industry standard behind Meta's broader "AI info" labeling on organic posts. The practical takeaway: the assumption that undisclosed third-party AI creative will quietly pass is no longer safe, and building a strategy around evading detection is building on sand.
Platform labels are only one of the two systems, and it is the less dangerous one. The sharper edge is the Federal Trade Commission's authority over deceptive advertising, which applies to every ad regardless of what any platform does or does not label. The centerpiece for AI UGC is the FTC's final rule on the use of consumer reviews and testimonials, which took effect on October 21, 2024. It prohibits creating, purchasing, or disseminating fake or AI-generated consumer reviews and testimonials — specifically those that misrepresent the identity, experience, or existence of the person supposedly giving them. Civil penalties can reach tens of thousands of dollars per violation, and each fabricated piece can count.
Read that against what an AI UGC ad is. A synthetic actor, generated to look like an ordinary person, delivering a scripted first-person account of using and loving a product they never used — that is, on its face, a testimonial that misrepresents the existence and experience of the reviewer. It does not matter that the creative was cheap or that everyone in the industry is doing it. Alongside the fake-review rule sits the long-standing FTC Endorsement Guides, which require that endorsements reflect the honest opinions and real experience of a genuine endorser, and that any material connection between the advertiser and the endorser be disclosed clearly and conspicuously. A generated persona has no honest experience to reflect and no independent relationship to disclose, which is precisely why staging it as an independent customer is the trap. The regulatory backdrop here connects to the broader questions in [the AI influencer manipulation trend](/guides/ai-influencer-manipulation-trend) and to how jurisdictions are starting to police humanlike synthetic personas in [regulating humanlike AI](/guides/humanlike-ai-regulation-avatar-persona-content).
Not all AI advertising carries the same risk, and understanding why sorts the safe uses from the dangerous ones. A glossy, obviously-produced AI brand video — a stylized scene, a clearly art-directed spokesperson, a product beauty shot — sits at low deception risk, because no viewer mistakes it for a candid clip from a real customer. It looks like an ad, it is understood as an ad, and an "AI info" label on it is a transparency nicety rather than a rescue from misrepresentation. The best-practice playbook for that polished lane is [AI UGC ads best practices](/guides/ai-ugc-ads-best-practices) and, for Meta's own generative ad tools, [Meta AI multimedia ads best practices](/guides/meta-ai-multimedia-ads-best-practices).
The UGC-style format is the opposite by design. Its entire conversion advantage comes from not looking like an ad — from reading as a real person, filming themselves, sharing an unpaid, unscripted, genuine opinion. That impression is the product. And it is exactly the impression the FTC's testimonial rules exist to protect, because a purported real-customer testimonial carries persuasive weight a labeled ad does not. So the same synthetic-actor technique flips from acceptable to actionable based purely on framing: a declared brand persona presenting your message is fine; that identical avatar staged as "just a customer who found this and had to share" is a fabricated testimonial. The technology is neutral; the framing is what the law reads. This is why the format cannot be governed by a single toggle — the risk lives in the presentation, not the pixels.
Hold the two systems apart deliberately, because passing one does not pass the other, and most compliance failures come from treating them as one gate. Disclosure is satisfied by labeling: the "AI info" marker is present, the political self-certification is completed where required, and you have not tried to strip provenance metadata to dodge detection. Deception is satisfied by honesty: the claims are true and substantiated, material connections are disclosed, and no fabricated person is presented as a genuine independent customer. You can pass disclosure and fail deception — a perfectly labeled AI clip that still stages a fake testimonial. You can also, in theory, be honest in substance while under-disclosing the AI, which is the failure the platform labels are closing. A compliant AI UGC ad has to clear both gates, and they are checked by different authorities with different remedies.
The practical version of this is a two-column check on every creative before it ships. Column one, disclosure: is the AI media labeled or detectable-and-labelable, and if the ad touches social, electoral, or political themes, have you completed Meta's self-certification? Column two, honesty: is every claim true, is the persona framed as what it actually is rather than as an unpaid stranger, and would a reasonable viewer be misled about who is speaking and why? A yes on column one and a no on column two is the exact profile of the ad that gets labeled, runs, converts, and then draws an enforcement action — the worst outcome, because it looked compliant the whole time.
None of this makes AI UGC unusable. It makes the reckless version unusable and rewards the disciplined one, which is generally how a format matures from a loophole into a channel. The durable pattern has a few consistent moves. Keep the persona openly a brand persona — a declared, recurring spokesperson your audience comes to recognize as representing you — rather than a rotating cast of fake strangers each pretending to be an unaffiliated customer. A recognizable, honest AI presenter is an asset that compounds; a disposable counterfeit customer is a liability that accrues. Make the claims independently true, because AI does not lower the substantiation bar; a synthetic mouth saying an unsupported claim is still an unsupported claim. And treat the platform label as a feature, not a wound — audiences in 2026 increasingly expect AI in advertising, and the reputational cost of being caught hiding it exceeds the cost of the small "AI info" marker.
The operational challenge underneath the compliance one is volume with consistency. The reason AI UGC took over is that a performance operation runs dozens of creative variants a week, kills the losers, and scales the winners — and doing that by hand across the persona, the script, the captions, the aspect ratios, and the platforms does not scale. The temptation is to solve throughput by spinning up an anonymous fake-customer factory, which is the precise move that trips the FTC line. The better solve is a system that produces high volume from a consistent, declared brand identity, so the output stays recognizable and honest as it scales instead of fragmenting into a hundred disposable counterfeits. That is a production-system problem, and it is where the choice of tooling quietly determines whether your compliance posture holds. The concrete, step-by-step version of the labeling workflow lives in [how to disclose AI-generated ads](/how-to/disclose-ai-generated-ads).
[Kompozy](/) fits this problem because its whole model of AI video is the compliant one by construction: a declared, recurring brand persona rather than an anonymous fake customer. It is a full AI generation-and-publishing engine, and its persona video formats — [Persona Shorts](/glossary/persona-shorts), Persona HeyGen, Persona VFX, Persona Frames, and Marketing Shorts — render an avatar that is openly a brand spokesperson drawn from your AI Influencer persona pool, not a synthetic stranger staged as an unpaid reviewer. That distinction is exactly the one the FTC's testimonial rule turns on. When your AI UGC-style creative is transparently a brand persona delivering your message, you are in the "declared spokesperson" lane the law tolerates, not the "counterfeit customer" lane it prohibits — the safest posture is baked into the format instead of bolted on afterward.
The honesty gate has a home in the workflow too. The [Persona Brief](/glossary/persona-brief) governs what the persona says — voice, allowed claims, and banned-word rules — on every generation, so an unsupported or off-limits claim is filtered at the point of production rather than caught in review, and the same brief keeps the identity consistent across a hundred variants so scale does not fragment into a hundred different fake people. Gemini face-lock holds one face across every clip and HyperFrames renders pixel-exact brand styling, which is the technical form of "recognizable, declared brand" — the opposite of the disposable-stranger pattern that draws enforcement. Producing fifty on-brand variants from one governed identity is the volume performance advertising wants without the anonymous-testimonial risk that volume usually creates.
Where Kompozy stops is worth stating plainly, because the disclosure step itself is not something a generation engine can complete for you. Kompozy produces the on-brand creative and publishes it across its supported platforms with scheduling and a per-post review gate, but the platform-side actions — completing Meta's AI self-certification when an ad falls under the social, electoral, or political rules, keeping the "AI info" label rather than evading it, and making the final legal call on whether a given claim is substantiated — happen in Ads Manager and in your own compliance judgment. The engine gives you the compliant-by-design format and the consistency to run it at scale; you still own the disclosure toggle and the honesty of the claim. Used that way, AI UGC is a channel you can build on rather than a bet against enforcement.
AI-generated UGC is the performance-creative format of 2026, and Meta's clearer AI ad labels are the market and the regulator catching up to it. The rules are not one thing but two, and the whole game is keeping them apart. Disclosure — Meta's "AI info" label, the political-ad self-certification, and the automated third-party detection that expands on June 1, 2026 — is a labeling task, and the honest move is to let the label stand rather than dodge it. Deception — the FTC's 2024 ban on fake and AI-generated testimonials and its standing endorsement rules — is a legal line, and no amount of labeling cures a fabricated person staged as a real customer. The format survives for the operations that treat both gates as real: a declared, recognizable brand persona instead of a counterfeit stranger, claims that are actually true, and disclosure treated as a feature. Run it that way and AI UGC is a durable channel. Run it as a fake-customer factory and the label you skipped is the least of the problems coming.
It depends on the ad and the edit. For ads about social issues, elections, or politics, Meta requires advertisers to self-disclose photorealistic images, video, or audio that were created or altered with AI to depict a real person doing or saying something they did not, or to fabricate a realistic person or event. For ordinary commercial ads there is no manual self-disclosure toggle for every creative, but Meta applies an "AI info" label to content made with its own generative tools, and from June 1, 2026 uses automated detection to label third-party AI-generated media in ads. Separately, the FTC's truth-in-advertising and fake-testimonial rules apply to any ad regardless of platform labeling.
Disclosure is a labeling question — telling the viewer the media was generated or edited with AI, which Meta handles with its "AI info" label and its political-ad self-certification. Deception is a legal question — presenting a fabricated person as a real, unpaid, genuine customer. An AI UGC ad can be fully labeled as AI-made and still be an illegal fake testimonial if the synthetic actor is framed as a real customer sharing a real experience. Labeling the media does not cure a fabricated endorsement; the FTC's 2024 rule bans fake and AI-generated testimonials outright.
When Meta detects AI-generated or AI-edited media in an ad — or the ad was made with Meta's own generative creative features — it attaches an "AI info" marker that appears on the "About this ad" screen in the three-dot menu, and sometimes next to the Sponsored label. Significant edits like background generation or generating the visual subject trigger it; minor edits like cropping, resizing, or color correction do not. Advertisers cannot remove an automatically applied label. From June 1, 2026 Meta also runs automated detection for third-party AI tools, labeling that media without any advertiser action.
The FTC's final rule on consumer reviews and testimonials, effective October 21, 2024, prohibits creating, buying, or disseminating fake or AI-generated reviews and testimonials that misrepresent the identity, experience, or existence of the reviewer. For AI UGC that means a synthetic actor cannot be presented as a real customer describing a genuine experience they never had. It also carries the standing FTC endorsement rules: material connections must be disclosed clearly and conspicuously, and endorsements must reflect honest, real experience. Penalties can reach tens of thousands of dollars per violation.
Because the format's entire power comes from looking unproduced and genuine — a real person, filming themselves, giving an honest, unpaid opinion. That is exactly the impression the FTC's testimonial rules exist to protect. Polished, obviously-produced AI ads carry a lower deception risk because no one mistakes them for a real customer's candid clip. An AI UGC ad is engineered to create that mistake, which is why the same synthetic-actor technique that is fine for a stylized brand video becomes a legal problem the moment it impersonates a genuine customer testimonial.
Keep the persona openly a brand persona rather than a fabricated stranger, make any claims true and substantiated, disclose material connections, do not stage a synthetic actor as a real independent customer, complete Meta's AI self-certification where the ad falls under the social/political rules, and keep the platform "AI info" label rather than trying to evade detection. The safe frame is a declared brand spokesperson or clearly-stylized creative, not a counterfeit of an unpaid customer. Treat disclosure and honest claims as two separate gates you have to pass, not one.
AI-generated UGC ads — synthetic actors filmed to look like a real customer talking to their phone — became 2026's dominant performance format, and Meta's clearer AI ad labeling is the response. Meta applies an "AI info" label to detected or Meta-made AI ad media, requires self-disclosure on social/electoral/political ads that use AI to depict real people or fabricate realistic events, and from June 1, 2026 auto-detects third-party AI media in ads. Layered on top, the FTC's 2024 rule bans fake and AI-generated testimonials outright. The key distinction: disclosure is a labeling task; deception — passing off a fabricated person as a genuine customer — is a legal line, and labeling the media does not cure it.
Get started → · ← All guides · Compare Kompozy vs other tools