An AI-generated UGC ad can look like a creator recommendation even when no customer recorded it. That difference matters. Before a paid ad is exported, the team should be able to answer three separate questions: Is the message truthful? Are the person, voice, product assets, and other media authorized? Does the final ad carry every disclosure required by the target jurisdiction and platform?
This checklist is for marketing, creative, and review workflows. It is not legal advice and does not determine whether a particular campaign complies with law. Rules vary by location, product category, audience, platform, and the facts of the ad. Use counsel for campaign-specific advice.
Review three layers before publishing
Do not reduce compliance to a single "AI-generated" badge. Review these layers independently:
| Layer | Questions to resolve | Evidence to retain |
|---|---|---|
| Law and regulator guidance | Is the ad deceptive? Is there an endorsement, testimonial, material connection, deepfake, regulated claim, or jurisdiction-specific AI disclosure duty? | Legal review, claim substantiation, disclosure decision, target markets, audience and category notes |
| Platform policy | Does the destination require an AIGC label, ad disclaimer, identity permission, commercial-content disclosure, or another format-specific control? | Policy version and date, account setting, screenshot or export, submission ID and review result |
| Contracts and rights | Who owns or licenses the source images, footage, voice, music, copy, trademarks, likeness, and final commercial use? | Licenses, releases, creator agreement, tool terms, invoices, source manifest and approval history |
Passing one layer does not prove the others. An ad can receive platform approval and still contain a false testimonial. A paid-relationship disclosure does not grant music or likeness rights. A license to use an image does not substantiate the product claim written over it.
Separate a dramatization from a testimonial
Start by classifying what the person or avatar appears to communicate.
Fictional demonstration
An obvious fictional demonstration can show a product workflow without claiming that the actor is a real customer. Keep the language observational and verifiable. For example, show the product feature and describe what is visible instead of scripting a personal experience that never happened.
Endorsement or testimonial
If reasonable viewers could understand the ad as a person's opinion, finding, belief, or experience, review it as an endorsement or testimonial. Do not invent a customer, purchase, result, or quote. Do not give an AI avatar a line such as "I used this for thirty days" unless the underlying experience is real, authorized, accurately represented, and supported by records.
The FTC's Consumer Reviews and Testimonials Rule Q&A says there is no blanket prohibition on AI stock avatars. It also explains that fake or false testimonials remain prohibited and that using an actor to portray a testimonialist can still be deceptive. The FTC's 2023 Endorsement Guides update expressly addresses virtual influencers, fake reviews, clear and conspicuous disclosures, and advertiser responsibility.
Use a stricter internal rule than "the script never says customer." Ask what the entire ad implies through the avatar, setting, captions, product handling, before-and-after sequence, star rating, quote styling, and CTA.
How regulations apply to AI-generated UGC in paid ads
There is no universal disclosure text that clears every AI-generated ad. Build a market-by-market decision record.
United States: deception, endorsements, and material connections
For U.S.-facing advertising, the FTC requires truthful advertising and treats a material connection as information viewers may need to evaluate an endorsement. The FTC's Disclosures 101 guidance advises placing an understandable disclosure with the endorsement where it is hard to miss. For video, the disclosure should appear in the video rather than only in a description; visual and audio disclosure may both be appropriate for how people watch.
The FTC's rule also prohibits specified fake or false consumer and celebrity testimonials, including a testimonial that misrepresents a nonexistent person's experience. An AI label does not cure a fabricated product experience or unsupported claim.
European Union: AI Act transparency
Article 50 of the EU AI Act applies from August 2, 2026. The European Commission's Article 50 transparency guidance explains obligations for certain generative AI systems and deepfakes. The consolidated AI Act text requires deployers of systems that generate or manipulate deepfake image, audio, or video content to disclose that the content was artificially generated or manipulated, subject to the regulation's scope and exceptions.
Do not assume that every AI edit is a deepfake or that one label format fits every EU placement. Record why Article 50 is or is not triggered, how the disclosure is presented, and whether other consumer, privacy, personality, copyright, sector, or national rules also apply.
Other jurisdictions and regulated categories
Map where the ad will be shown, not just where the creative team sits. Check the current rules for each target market and for categories such as health, finance, employment, housing, alcohol, gambling, political advertising, and products directed to children. A general-purpose checklist cannot replace category-specific review.
Apply the live platform policy
Platform rules can be narrower, broader, or more operational than a legal requirement. Check the live policy again when the ad is submitted.
TikTok Ads
TikTok's Misleading and false content policy was updated in April 2026. It requires an AIGC label or another clear disclosure for completely AI-generated or significantly AI-modified ad media and says an undisclosed ad may be rejected or restricted. The same policy prohibits unauthorized misuse of a person's likeness, including misleading public-figure endorsements.
TikTok's ad disclaimer guidance identifies AI-generated, synthetic, or manipulated media as a mandatory disclaimer use case. Its intellectual property advertising policy also prohibits unauthorized third-party media, brands, names, logos, and other infringing ad content.
Treat the account toggle as an execution control, not the whole review. The exported creative, caption, landing page, identity permission, claims, and commercial-content disclosures still need their own checks.
Google Ads
Google's July 2026 AI labeling update permits advertisers to add labels directly to AI-generated or AI-edited image and video creatives or use Google's AI label setting as it becomes available across its advertising products. Google also states that using the setting does not guarantee compliance with a specific regulation.
Google's AI transparency explanation says disclosure duties vary by region and content type. Record the campaign market, format, asset treatment, label method, and any additional policy that applies. Do not infer that an unlabeled format is lawful merely because a particular account does not show the label control.
Clear every source and identity right
Create a source manifest before generation. For each input and output, record the file, owner, license, permitted use, territory, duration, modification rights, sublicensing or platform-use terms, and any attribution requirement.
Person, avatar, voice, and likeness
- Identify every real person represented in the source or final ad.
- Obtain a release covering commercial advertising, AI editing, voice or likeness synthesis, target platforms, territories, duration, revisions, and withdrawal or expiration terms.
- Confirm that the person authorized the actual script, product, claim, and context. Permission to create an avatar is not automatically permission to endorse every product.
- Do not use a public figure, creator, employee, or private person to imply sponsorship, use, or approval they did not give.
- Keep the approved version and a record of later edits. A new claim, language, product, or market may require renewed approval.
Product and brand assets
- Confirm rights to product photos, packaging, logos, screenshots, reviews, quotes, and marketplace listing content.
- Check whether a supplier or marketplace license permits off-platform paid advertising and AI modification.
- Avoid presenting a third-party trademark in a way that implies affiliation or endorsement.
- Preserve the source page or document that supports each price, feature, comparison, result, safety, environmental, or performance claim.
Music, footage, fonts, and generated output
- Verify that music and stock-media licenses cover the specific paid-ad platform, geography, duration, spend tier, and editing method.
- Check the AI provider's terms that applied when the asset was created. Record commercial-use limits, restricted content, indemnity terms, and any output-ownership language without assuming the tool transfers third-party rights.
- Review output for recognizable people, copyrighted characters, logos, watermarks, protected designs, or copied source elements.
- Keep provenance and generation records, but do not treat provenance alone as proof of ownership or permission.
Substantiate the advertising message
Review the script, visuals, audio, captions, overlays, landing page, and implied sequence as one claim set.
| Review item | Approval question |
|---|---|
| Personal experience | Did the represented person actually have the experience, or is the scene clearly a dramatization? |
| Product performance | Is every objective claim supported before publication by evidence appropriate to the claim? |
| Demonstration | Does the shown result come from the advertised product under the conditions viewers are likely to understand? |
| Comparison | Are the compared products, versions, dates, prices, and test conditions current and fair? |
| Typicality | Does the ad imply a result is typical, and if so, is that implication supported or appropriately qualified? |
| Offer | Are price, subscription, renewal, eligibility, shipping, inventory, and material limitations accurate on the ad and destination? |
| Landing page | Does the destination preserve the same product, offer, claims, disclosures, and advertiser identity? |
Avoid unsupported "best," "number one," guaranteed-result, instant-result, medical, financial, or environmental claims. A disclaimer cannot contradict the main message, and a disclosure should not be hidden behind a click, at the end of a video, or among unrelated hashtags.
Use a documented preflight workflow
Before handing the file to the media buyer, require one review record:
- Identify markets, audience, product category, platforms, placements, languages, and campaign dates.
- Classify each synthetic element: script, image, person, voice, scene, product demonstration, edit, or background.
- Decide whether the ad communicates a dramatization, endorsement, testimonial, or real customer experience.
- Attach source licenses, likeness and voice releases, tool terms, claim evidence, and creator or agency agreements.
- Select the legal and platform disclosures, then place them where viewers will notice and understand them.
- Review the full ad and landing page for express and implied claims.
- Save the approved master, disclosure text, platform settings, approver, timestamp, and policy links.
- After submission, record platform review results and preserve the exact served version.
- Re-review before changing the script, avatar, offer, language, market, platform, or campaign context.
A practical stop rule is simple: if the team cannot identify the person represented, the rights holder, the source of a claim, the required disclosure, or the approving owner, do not publish the ad.
Build a reviewable creator-style ad in Reelsy
Reelsy's UGC Ad Generator prepares a creator-style brief, script, and storyboard for review. It does not turn an AI avatar into a real customer, supply legal clearance, submit the ad to a platform, or prove a performance claim. For multiple reviewable product-ad directions, use the Video Ad Variation Generator. Use the AI UGC ads ROI guide to define measurement after the rights, claims, and disclosure review is complete.
Before production, review Reelsy's Privacy Policy, Terms of Service, and current pricing and credits. When the brief, source permissions, approved claims, and disclosure plan are ready, open the product-ad workflow in Content Studio. Human approval, legal review, export, platform submission, campaign setup, and ongoing compliance remain separate responsibilities.
AI-generated UGC ads disclosure FAQ
Do AI-generated UGC ads always need an AI label?
There is no single worldwide rule for every ad. The answer depends on the audience location, the content, the platform, and whether the asset realistically depicts a person, event, or endorsement. TikTok currently requires disclosure for completely AI-generated or significantly AI-modified ad media, while Google offers AI labels and notes that using its label does not itself guarantee legal compliance.
Can an AI avatar give a customer testimonial?
Do not present an AI avatar as a real customer or imply an experience that did not occur. The FTC says there is no blanket ban on AI stock avatars, but fake or false testimonials and misleading portrayals can violate the Consumer Reviews and Testimonials Rule or the FTC Act. Use an obvious fictional demonstration or a truthful, authorized testimonial with records supporting the underlying experience.
Can an advertiser clone a person's face or voice for an AI UGC ad?
Only proceed after confirming documented permission that covers the specific person, media, editing, voice or likeness use, commercial advertising, platforms, territories, duration, and permitted revisions. Platform approval is not a substitute for that permission. TikTok's advertising policy prohibits unauthorized likeness misuse, particularly misleading public-figure endorsements.
Is a platform disclosure tool enough for an AI UGC ad?
Not necessarily. A platform label may address one platform requirement, but it may not disclose a paid relationship, prevent a false testimonial, substantiate a product claim, or satisfy another jurisdiction's rules. The FTC advises that a built-in disclosure tool may not make a material connection clear and conspicuous by itself.
What rights records should an AI UGC ad team keep?
Keep the brief, source-file licenses, creator or performer agreement, voice and likeness release, music and font licenses, trademark permissions, AI tool terms in effect at creation, output provenance, claim evidence, disclosure decision, approvals, final export, platform submission, and any later changes.
Is this AI UGC advertising checklist legal advice?
No. It is a production and review checklist, not legal advice. Laws, self-regulatory codes, contracts, platform policies, product categories, and disclosure requirements vary by jurisdiction and change over time. Ask qualified counsel to review the actual campaign when the legal position matters.
The FTC, European Commission, EU AI Act, TikTok Ads, and Google Ads sources linked above were checked on October 15, 2026. Policies, laws, product interfaces, and enforcement practices can change. Recheck the live rules and obtain campaign-specific advice before publishing.