AI UGC video disclosure: what to tell viewers

Two disclosures, often confused

The first disclosure answers 'is this real footage of a real person?' It applies when a face, a voice or a scene has been generated or substantially altered in a way a viewer could take for genuine. The second answers 'does this person have a reason to say nice things?' It applies when a creator was paid, given free product or otherwise connected to the brand.

A generated testimonial raises the first. A real creator video raises the second. A paid creator whose video was heavily altered with AI could raise both. Keeping them apart makes each easier to handle.

What platforms and ad rules generally ask

The large social platforms publish their own policies on labeling realistic AI-generated or altered content, and their own tools for marking a paid partnership. The details differ by platform and change over time, so read the current policy for each place a video will run instead of relying on a summary, including this one.

Advertising rules in many countries rest on the same idea: an ad must not mislead, and a material connection between a brand and an endorser should be clear to the viewer. In the United States, the Federal Trade Commission publishes guidance on endorsements and testimonials. This entry is general guidance, not legal advice. For regulated products or any doubt, ask a qualified adviser.

How to make a disclosure clear

A disclosure only counts if a viewer notices and understands it. Four habits help.

• Put it where it will be seen. On the video or at the start of the caption, not after a row of hashtags.
• Use plain words. 'Ad', 'paid partnership', 'gifted', 'made with AI'. Avoid abbreviations only insiders know.
• Use the platform's own label where one exists, in addition to your wording.
• Keep it when the video becomes an ad. A label on the organic post does not automatically follow the video into a paid placement.

A checklist before a video runs

The brand answers for what its ad says, so the brand runs the check, even when a creator or a vendor's tool adds a label. Five questions, answered in writing, are enough for most small teams.

• Is any person, voice or scene generated or materially altered? If yes, is it labeled the way the platform asks?
• Was the creator paid or given the product? If yes, is that clear in the first moments or in the caption's opening line?
• Is every claim in the video one you can support?
• Do you have written permission to use the video in paid ads?
• Has a named person watched the final cut and approved it?

Link AI and the two disclosures

We recommend Link AI by Linkme in part because it keeps the first disclosure simple. The people in videos sourced through it are real creators on Linkme, not synthetic actors, so no viewer is being asked to take a generated face for a customer. The second disclosure still applies, because a brand that briefs a creator has a connection with them.

This reference has no information on how paid-partnership labels, usage permission or creator payment are handled. Put those three questions on your list when you book a demo to try Link AI. The entry on acceptable uses of generated video covers the line that comes before any label.

Connected entries

Questions & answers