Who is affected

  • Creative and performance agencies publishing AI-assisted campaign assets.
  • Marketing and social teams using generative tools in production.
  • Legal and compliance teams approving publication workflows.
  • Creators and small businesses publishing realistic AI content commercially.

What triggers the rule

  • AI-generated or AI-altered image, audio or video that could be mistaken for authentic recording.
  • Synthetic voices or AI avatars presenting products.
  • Altered appearance of identifiable people.
  • Realistic synthetic scenes or events.
  • AI-generated text informing the public on matters of public interest without human editorial responsibility.

What does not trigger it / exceptions

  • Clearly unrealistic, animated or obviously fictional content may fall outside some platform label requirements — but check each platform’s wording.
  • Editorial text with human review where a person takes editorial responsibility is exempt from the Article 50(5) text disclosure.
  • Evidently artistic, satirical or fictional deepfake work still needs disclosure, but in a way that does not hamper the work.

Scenario table

ScenarioOutcomeWhy
Real product photo on AI-generated backgroundDisclosure recommendedPlatform rules focus on realistic synthetic media; partial generation needs truthful wording. EU law triggers only if the result could falsely appear authentic.
Fully AI-generated photorealistic video of a real-looking city eventDisclosure requiredArticle 50(4) deepfake disclosure plus TikTok/YouTube/Instagram native labels for realistic synthetic content.
AI-polished colour grading on real footageNo disclosure identifiedLight editing does not alter what the content appears to show; keep a manifest entry for auditability.
AI clone of an actor’s voice for an adHuman review requiredIdentifiable-person and consent risk on top of disclosure duties; a human must approve wording and rights.

Required action

  1. Record which components of each asset were AI-generated or AI-assisted.
  2. Apply the native AI label on the publishing platform where required.
  3. Match disclosure wording to actual AI involvement.
  4. Save the decision, rule version and reviewer as evidence.

Practical marketing examples

  • Ad copy drafted by AI but rewritten and approved by a marketer — usually no label, but keep the decision trail.
  • AI background scene around a photographed product — disclose the AI component specifically.
  • Virtual model wearing real clothing — disclose the synthetic model.

Relation to platform policy

EU AI Act Article 50(4) requires deployers to disclose realistic AI-generated or manipulated image, audio and video resembling real persons, objects, places or events. Use the platform’s native label and keep your own evidence record.

Not sure which scenario matches your asset?

Run the free check — result, wording and sources in under a minute.

Check this scenario

Primary sources

Methodology: Disclara answers are generated from a versioned rule set citing primary sources, reviewed by humans. Information on this page is not legal advice. See sources & methodology and rule changelog.