AI disclosure scenario
Do you need to disclose a fully synthetic AI ad?
Direct answer
Usually yes when the ad contains realistic generated people, voices, products, places or events. Apply the destination platform's native label, add plain-language wording that accurately describes the generated components, and review every product or performance claim separately. Disclosure does not validate any marketing claim or likeness permission.
Applies when
- A fully synthetic ad is not automatically an Article 50 deep fake merely because AI created every pixel. The binding trigger is whether its image, audio or video resembles existing persons, objects, places, entities or events and would falsely appear authentic or truthful.
- When that test is met, the deployer must clearly and distinguishably disclose the artificial origin by first exposure. The label does not make a misleading product demonstration, fabricated testimonial or unauthorised likeness lawful.
Doesn't apply when
- A clearly fantastical or abstract animation that does not resemble real persons, places, products or events, unless a broader platform or client rule applies.
- AI used only for ideation, storyboarding or draft copy when none of the generated material appears in the final ad.
- Minor production assistance such as captions, repair or colour correction that does not change what the advertisement represents.
Worked decision
Example
A paid video ad uses a photoreal AI presenter, generated product demonstration, synthetic voice and generated location.
Decision: Disclosure required
Every audience-facing media component is synthetic and presented realistically. Use the native platform label, visible wording and human review of the product demonstration before trafficking.
This advertisement was generated with AI, including the presenter, voice and setting.
Next steps
Required action
- Record the generated or altered component, source asset, tool and destination before publishing.
- Apply the native platform label wherever the applicable platform workflow requires it.
- Use component-specific wording such as: “This advertisement was generated with AI, including the presenter, voice and setting.”
Publishing layer
Platform differences
TikTok
Label realistic AI-generated image, audio and video. Fully generated ads can also be removed if they violate impersonation, misinformation or other policies despite the label.
Open platform guide →YouTube
Select Yes under AI use for realistic generated or meaningfully altered content. YouTube can also apply labels from C2PA or its detection systems.
Open platform guide →Instagram and Facebook
Meta can label generated image, audio and video from technical signals or publisher disclosure. Advertising and claim review remain separate from AI provenance.
Open platform guide →Evidence
Source
Freshness and accountability
Last verified
Last reviewed 2026-09-12 · Current
Run this exact scenario
The content type, AI operation, context and starting platform for this scenario are preselected. Confirm the remaining facts.
Copy-ready language
Disclosure wording options
This advertisement was generated with AI, including the presenter, voice and setting.
AI-generated ad. Product packshot is real; presenter, voice and demonstration are synthetic.
Related guides
“Doesn't apply” means no disclosure requirement is identified for the facts stated here. A client policy, advertising rule or another destination platform can still be stricter. This page is source-backed information, not legal advice.