Classified examples · Customers and content · Po polsku
AI avatar or synthetic voice in training and promo videos
Transparency (Article 50)Deployer
The company produces videos with an AI presenter or an AI voice-over, sometimes based on a real employee.
How we classify it
- Deployer of an AI system generating video and audio.
- Not high-risk and not prohibited.
- A realistic synthetic person or voice is a deepfake under Article 50(4): disclose that the content is artificially generated or manipulated.
- Result: transparency obligations; consent is needed where a real person's face or voice is used.
What applies now
- Visible or audible disclosure, for example at the start and in the description
- Written consent from any employee whose likeness or voice is used, with the right to withdraw
- Keep the provider's machine-readable marking
What would change the answer
- Imitating a real person without consent can breach personality rights and GDPR regardless of the AI Act
Evidence to keep
- Consent forms
- Disclosure text
- Transparency checklist
In the AI Compliance Kit: 05Transparency Checklist and Disclosure Texts · 02AI Acceptable Use Policy.
Dates and sources
- Transparency obligations, Article 50: 2 August 2026
- Machine-readable marking of generative AI output: 2 December 2026
Legal basis: AI Act art. 50(4).
Similar examples
- A vendor's AI chatbot answers customers on our website Transparency (Article 50)
- An AI voice agent answers the phone Transparency (Article 50)
- Marketing creates images with an AI image generator Transparency (Article 50)
- AI writes blog posts, newsletters and product descriptions Depends on the use