In short: AI-generated “UGC” swaps the creator for a synthetic person. AI tools are not banned as such, but the rules around them are tightening: the US FTC’s rule on fake reviews names AI-generated reviews by people who do not exist, the EU AI Act’s transparency duties for deepfakes apply from 2 August 2026, and TikTok, YouTube and Meta each have labeling rules. When a video’s value is a real person’s real experience of a real product, a synthetic presenter cannot supply it. AI works well behind the camera, for scripts, outlines and captions.
What people mean by AI-generated UGC
UGC normally means a video filmed by a real person who has used the product. “AI UGC” usually means a video in which software produces the presenter, the voice or the whole scene. In practice there is a spectrum, and the rules treat its ends very differently:
- AI helps behind the scenes (a script outline, captions, translation, audio clean-up): normally no label. YouTube lists these as minor edits that need no disclosure.
- AI edits real footage (color and lighting filters, a tidier background): YouTube lists color and lighting filters as minor edits. Filters that change how skin or a product looks can still mislead in a beauty video, so keep footage true to life.
- AI generates the person or the scene (a synthetic presenter “trying” a product): labeling duties apply, and a made-up customer opinion raises the fake-review question below.
What the rules say
This is general information, and whether a particular ad is compliant is a legal question for the brand and its adviser.
United States: fake reviews and testimonials
The FTC’s final rule on fake reviews, announced in August 2024, covers reviews and testimonials that misrepresent their author. The FTC describes them as including fake reviews by nonexistent people, such as AI-generated fake reviews, and reviews by people who did not have actual experience with the business or its products. Businesses may not create or sell them, and may not buy or spread them when they knew or should have known they were fake or false. The FTC can seek civil penalties against knowing violators. A synthetic presenter who says “I have used this for three weeks” describes an experience nobody had, which is why brands should treat that format with great care. The FTC’s influencer guidance also tells endorsers to say only what they have tried (see UGC disclosure rules). For where paid UGC ends and a customer review begins, see paid UGC vs customer reviews.
European Union: the AI Act’s transparency rules
The European Commission states that the transparency rules in Article 50 of the AI Act apply from 2 August 2026. Providers of generative AI have to mark synthetic content in a machine-readable way, and people who use AI systems (“deployers”) have to disclose deepfakes. The Commission defines a deepfake as AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear authentic or truthful. Its page also mentions a grace period until December 2026 for the marking duty of generative AI systems that were already on the market before 2 August 2026, and lists fines of up to €15 million or 3% of worldwide annual turnover for companies. Whether a given video counts as a deepfake is for the Commission’s guidelines and the national authorities to say.
The platforms
- TikTok: Synthetic or manipulated media that shows realistic scenes must be clearly disclosed, for example with a sticker or caption such as “synthetic”, “fake”, “not real” or “altered”. Realistic synthetic media that is not prominently disclosed is not allowed, and synthetic media of public figures may not be used for endorsements.
- YouTube: Creators have to disclose realistic content that is AI-generated or meaningfully altered, for example a realistic scene that did not occur or a real person made to appear to say or do something they did not. Minor edits such as beauty filters, color or lighting filters, captions, audio repair and using generative AI for a video outline or script do not need a disclosure.
- Meta (Facebook, Instagram, Threads): Labels content as “AI info” when it detects industry-standard AI indicators or when people disclose that they are uploading AI-generated content, and may add a more prominent label when content carries a high risk of materially deceiving the public.
Rules differ by platform and change often. Read the current page for every platform you publish on.
Where real creators still win
This part is our view, not a source. A synthetic person cannot truthfully show how a foundation sits on real skin, how a shade looks in daylight, or how a jacket fits a body, because none of that has happened. Honest texture, fit and everyday-use footage is exactly what makes UGC convincing. Claims are a second problem: an AI presenter adds no evidence, and a cosmetic claim still has to be substantiated before it is made (see skincare and cosmetic claims in UGC).
What AI is good for
- Planning: outlines, hook ideas and script drafts that the creator then rewrites in their own words.
- Finishing: captions, subtitles, translation and audio repair.
- Briefs: a first draft of a brief that a person checks. A template is in the UGC brief for beauty and skincare.
Checklist for brands
- Ask in the brief for videos filmed by the creator, with no synthetic people presented as real.
- Do not publish a made-up customer opinion as if a real customer said it.
- If any AI-generated or heavily altered footage is used, label it the way each platform requires.
- Keep the claims list and the disclosure line in the brief.
Checklist for creators
- Film it yourself and say only what you really experienced.
- Use AI for planning and finishing, not to invent an experience.
- Follow each platform’s AI label rule when you post, and label paid content as an ad.
- Do not let a brand clone your voice or likeness without a written agreement.
On Usicri
Under the Terms of Service, every video must have been filmed by the creator and must show real people. AI-generated or synthetic persons may not be presented as real, and fake reviews or testimonials are not allowed (sections 3 and 8.1). On every upload the creator confirms this by ticking a declaration. The license a brand receives does not allow it to create synthetic replicas of the creator’s voice or likeness or to use the video to train AI models (section 5.2), and Usicri does not use videos to train AI models either (section 5.3). Usicri does not check videos for compliance with advertising law, so the checklists above are for you. For labeling paid posts see UGC disclosure rules.
The checks for a real creator, in practice, are in how to tell if a UGC creator is real.