ASA on AI deepfake ads: you answer for the output you run
CAP says a realistic AI celebrity endorsement can breach the Code and the advertiser stays responsible. Which Sume avatar inputs involve a real person's photo.
The UK advertising rules treat an AI deepfake endorsement like any other ad claim. CAP's news piece says that if an AI-generated celebrity "endorsement" looks real enough to mislead consumers, it is likely breaching the Code, and that the advertiser remains responsible for any output used in its ads.
The quotes are from the CAP News article dated 11 Jun 2026, read 2026-09-30. Sume's avatar facts are from the Create new avatar and Face swap docs. This is a summary, not legal advice.
What does CAP say about deepfakes and endorsements?
Four points on the page matter for an avatar ad.
| Point | As the page puts it |
|---|---|
| Media-neutral | The Code applies regardless of how the content was created |
| Celebrity endorsement | Likely breaching the Code if it looks real enough to mislead consumers |
| Personality rights | Personality rights and IP "aren't issues the ASA regulates", but take legal advice on using a public figure's likeness |
| Responsibility | The advertiser remains responsible for any output used in their ads, including from automated platforms |
Which Sume avatar inputs involve a real person's image?
Only some of them. An avatar can be made three ways: a **Prompt** describing the avatar, a **Profile** of structured traits, or an **Image**, a reference image. The first two start from text. The image route takes input.image_url, which must be a fetchable public HTTPS image URL; Sume rejects localhost, private-network, non-HTTPS and non-image URLs before generation.
Face swap is a Beta endpoint that applies a ready avatar handle to a public HTTPS source video, so the source video's person is the face being replaced.
Does Sume check consent or likeness?
The docs I read describe no likeness, celebrity or consent check. The image check is only that the URL is a fetchable HTTPS image. So the question CAP raises, whether the result could mislead consumers about a real person, stays with you before the ad runs.
What should I do before running an avatar ad?
Use a prompt or profile avatar when you do not need a specific person, and only upload a photo you have rights to. Review the finished video the way CAP describes: is it socially responsible, free of misleading claims, and right for the audience? Keep the job id and artifact for each avatar video as your record. For platform-side rules, see YouTube's likeness detection and AI avatars.
Sources
Related posts
More in Use cases
- Black Friday countdown video ad: burn the days-left text
Meta says countdown overlays create Black Friday urgency. A countdown is authored text, so burn a new days-left cue onto one clip each day with Sume.
- Bluesky video captions: WebVTT sidecar vs burned-in
Bluesky video embeds take WebVTT caption files next to the video. Sume burns captions into pixels, so build the VTT yourself from sentence segments.
- Break down a UGC ad by scene: $0.30 analysis or free stills?
To study a UGC-style ad, video inspect gives free stills and a $0.01 per minute transcript; the legacy analysis adds typed scenes for $0.30. What each returns.
- Virtual staging disclosure: California AB 723 and originals
California AB 723 asks listings with digitally altered photos to say so and link to the unaltered originals. How to keep original and edit as two URLs.
Written by Sume