AI UGC Ad Disclosure Checklist: Review Before You Publish

Editorial owner: Genflow Editorial · report a factual correction
Policy and source review completed: September 6, 2026

Before publishing an AI UGC ad, identify who or what the presenter represents, classify every spoken or implied product statement, attach proof and permissions, decide which disclosures must appear in the creative and the platform, and save the final label state with the approved file. A platform’s AI toggle is one control; it does not make a false testimonial, unsupported claim, or unauthorized likeness safe.

Use the checklist below as a production handoff, not as legal advice or a promise of platform approval. Laws, platform policies, account controls, formats, and regional requirements change. Have the responsible legal or policy owner review the real ad and recheck the destination immediately before launch.

Start with an Authenticity and Disclosure Card

Create one card for every final ad—not for a campaign folder full of unfinished variants.

Ad ID and final file checksum:
Brand / product / offer:
Destination, account, region, and placement:
Presenter type:
Person or likeness source:
Permission record:
Voice source and permission:
Script owner and approved version:
Product statements and evidence IDs:
Testimonial or endorsement status:
AI generation or material edit summary:
Creative-side disclosure:
Platform-side disclosure setting:
Final reviewer and review date:
Policy pages rechecked on:
Release receipt / campaign ID:

The card forces five questions into the open: who is depicted, what the ad asks viewers to believe, what proves it, what must be disclosed, and whether the published settings match the approved plan.

Gate 1: map the presenter before reviewing the script

“UGC style” describes a creative format. It does not prove that a real customer recorded the experience.

Presenter type	What viewers may infer	Required production record	Hold when
Real customer	A purchaser is describing personal experience	identity, consent, original statement, any material connection	the experience or permission cannot be verified
Paid creator or actor	A person is performing an approved script	contract, likeness/voice permission, compensation relationship, script	the performance could be mistaken for an undisclosed customer testimonial
Employee or founder	Someone connected to the brand is speaking	role, approved claim source, relationship disclosure decision	the connection is hidden where it could affect credibility
Synthetic stock presenter	No real customer supplied the depicted experience	asset/tool source, rights, script provenance, synthetic-person disclosure decision	the script uses first-person experience the synthetic presenter never had
Composite or digitally altered real person	A real likeness or voice has been materially changed	source assets, scope of edit, permissions for likeness and voice	the edit exceeds permission or creates a statement/action the person did not authorize

The FTC’s current Consumer Reviews and Testimonials Rule Q&A says there is no blanket ban on AI stock avatars, but a synthetic avatar can still create a deceptive testimonial depending on the underlying message. The safest operational rule is not “AI avatar allowed.” It is “do not represent invented experience as a real person’s experience.”

Gate 2: classify every statement the ad makes

Review the transcript, captions, product demonstrations, on-screen labels, before/after sequence, landing page, and implied story. Put each material statement into one row.

Statement ID	Exact words or visual implication	Type	Evidence	Owner	Result
C01	“I use this every morning”	personal-experience testimonial	original customer record or none	claims owner	pass / rewrite / remove
C02	product shown producing a result	demonstration claim	repeatable test and product setup	product owner	pass / qualify / remove
C03	“works in minutes”	performance claim	substantiation matching the advertised use	claims owner	pass / qualify / remove
C04	old image followed by improved image	before/after implication	provenance, conditions, typicality review	legal / claims	pass / qualify / remove
C05	“the brand sent this to me”	material-connection statement	contract or gifting record	creator manager	pass / correct

Do not approve a line merely because it came from a prompt, creator brief, competitor ad, customer review, or language model. The FTC’s advertising FAQ states that endorsements must be honest and cannot make claims the advertiser could not substantiate directly.

If a synthetic presenter says “I tried this for seven days,” one of two things must be true: the line faithfully communicates a documented real testimonial with appropriate attribution and permission, or it must be rewritten as a product statement that does not invent personal experience. An “AI-generated” label does not cure the false experience.

Gate 3: separate three disclosure decisions

Teams often collapse different disclosures into one switch. Record them separately.

A. Material connection

Does a real speaker, creator, employee, affiliate, or reviewer have a relationship with the brand that viewers may not expect? If so, decide how to disclose that relationship clearly in the ad’s actual format. The FTC’s Disclosures 101 emphasizes that the relationship should be obvious; for visual formats, the disclosure must be noticeable in the content, not buried where viewers are unlikely to see it.

B. Synthetic or materially edited media

Was the presenter, voice, event, or product demonstration completely generated or significantly altered? Which creative-side label, metadata, account control, or platform disclosure applies to this destination?

TikTok’s current AI-generated content guidance requires labels for realistic AI-generated content and prohibits some impersonating or harmful depictions even when labeled. TikTok Ads Manager also documents an AI-generated-content disclaimer for completely generated or significantly modified ad media.

Meta’s June 2026 update says its ads transparency system can add “AI info” for its own generative tools and for detected third-party AI signals, with placement varying by the kind of edit and region. Do not assume automatic detection transfers responsibility or behaves identically in every account.

C. Claim qualification

Does the ad need a limitation, eligibility condition, typical-results statement, price term, or other qualification so the overall impression is accurate? This is not the same as labeling media as AI-generated.

For each decision, write the exact disclosure, location, duration, contrast/readability requirement, spoken-audio requirement if applicable, platform setting, and owner. “Add disclaimer later” is not a release state.

Gate 4: review disclosure in the final placement

Check the rendered ad, not just a script document.

• Is the disclosure present before the relevant claim or impression has already landed?
• Can a viewer read it at normal playback speed on the target device?
• Does it survive captions, mute controls, CTA bars, usernames, crops, and platform overlays?
• If the ad relies on audio, is material information also available to viewers watching without sound when needed?
• Is the wording understandable without clicking a profile, expanding a caption, or opening a separate page?
• Does the landing page preserve rather than contradict the ad’s qualifications?

Use the existing vertical video safe-zone checklist to test collisions and crops. Safe placement is necessary, but it cannot rescue vague wording.

The FTC’s Endorsement Guides Q&A repeatedly treats clarity and context as decisive. Google Ads’ misrepresentation policy similarly focuses on whether the ad withholds or obscures information people need to make an informed decision.

Gate 5: make the platform handoff verifiable

A clean creative file can still be launched with the wrong account setting. Give the media buyer a release packet.

Final asset ID and checksum
Approved transcript and captions
Presenter classification and permissions
Claim ledger and evidence links
Required creative-side disclosures
Required platform-side toggle / disclaimer
Destination, region, placement, and account
Screenshot or export of the submitted setting
Submission time and campaign/ad ID
Reviewer and next policy recheck date

For duplicated or localized ads, do not assume a prior toggle, disclosure, or permission automatically carries forward. Copy the card, re-evaluate the destination and language, then create a new release receipt.

A practical decision matrix

Situation	AI label decision	Testimonial decision	Additional action
Synthetic presenter reads factual product copy	Check platform/region rules for generated media	Avoid first-person customer experience unless sourced and accurately attributed	substantiate every product claim
Paid real creator gives genuine experience	Label AI only if media meets the platform threshold	disclose material connection clearly	retain original statement and permission
Real creator’s voice is cloned for a new script	Treat as materially altered media and recheck platform rules	do not imply the creator said or experienced something unapproved	obtain specific voice and script permission
AI recreation of a celebrity or private person	A label is not sufficient by itself	high impersonation and permission risk	hold unless rights and platform policy are affirmatively cleared
Generated product demo shows an untested result	AI disclosure does not validate the demonstration	the visual itself may make a performance claim	replace with verified footage or a substantiated depiction

The matrix is a triage tool, not a legal conclusion. Escalate sensitive categories, health or financial claims, minors, public figures, political content, regulated products, realistic impersonation, and disputed rights to the responsible specialist.

Using the card with Genflow

Genflow can be used to create and reuse AI image/video workflows, but the policy records in this guide are not claimed as native Genflow fields. Keep the Authenticity and Disclosure Card in the team’s system of record. Pass only approved source media, scripts, claims, and permissions into the creative workflow; then link the final Genflow output to the release receipt.

If the ad fails a gate, fix the source record or creative. Do not generate more variants to bury an unresolved truth, rights, or disclosure problem.

Research and preparation notes

Genflow Editorial examined ten current regulatory, platform, and industry references, translated them into a creator-style ad handoff, and designed this article’s five-gate card. Automated tools helped organize evidence, shape the draft, and create the conceptual cover illustration.

The cover is an original workflow metaphor. It is not a platform interface, legal approval, campaign screenshot, rejection result, or customer case. No approval rate, policy outcome, conversion lift, or Genflow compliance guarantee is claimed.

Publish the receipt with the responsibility

Before launch, make sure the person approving the ad can answer five questions from the card: who is depicted, what the ad asks viewers to believe, what proves it, what disclosures apply, and which settings were actually submitted. If one answer is missing, the ad is not release-ready.