EU AI Transparency Guidelines
Commission guidance on Article 50 transparency obligations for providers and deployers of certain AI systems, including user disclosure and AI-generated content marking/labelling.
Commission guidance on Article 50 transparency obligations for providers and deployers of certain AI systems, including user disclosure and AI-generated content marking/labelling.
Relevant to covered interactive AI systems, deepfakes, emotion recognition/biometric categorisation, and certain generated or manipulated content.
Supports consistent interpretation and enforcement of Article 50 by competent authorities.
Penalty context: Underlying AI Act penalty rules apply; organisations should assess exact exposure against the Act.
Commission guidance on Article 50 transparency obligations for providers and deployers of certain AI systems, including user disclosure and AI-generated content marking/labelling.
Label covered deepfake content in a clear manner when required.
Actor: deployers · Domain: transparency
Evidence: Publication workflow, labels and exception rationale.
Design interactive AI systems so users can recognise they are interacting with AI where required.
Actor: providers · Domain: transparency
Evidence: Interface disclosure, product requirement and test evidence.
Implement applicable machine-readable marking or detection-support measures for AI-generated or manipulated content.
Actor: providers · Domain: traceability
Evidence: Technical marking specification and conformance tests.
Cybatar's mapping is designed for AI governance and assurance work. Legal interpretation remains anchored to the current primary text and competent authority guidance.
Open European Commission source