Chapter 1 — Article 50: What and When Are the Transparency Duties?
Legal Scope and Requirements
Article 50 of the EU AI Act applies to “providers” (anyone placing AI systems on the EU market, even if outside the EU) and, in part, to “deployers” (those operationalizing AI). You must implement prominent, clear, and accessible transparency measures if your product or service:
- Lets natural persons interact with AI (e.g., chatbots, voice agents)
- Obligation: Inform users they’re interacting with AI, unless it’s obvious to a well-informed, observant, and circumspect user.
- Involves emotion recognition or biometric categorization
- Obligation: Inform all individuals affected, provide proper safeguards.
- Generates or manipulates audio, images, video, or public-interest text (“deepfakes” or synthetic content) that could be mistaken for authentic
- Obligation: Embed robust machine-readable labeling and visible/audible disclosure. Minor edits and closed industrial systems are generally excluded.
- Publishes potentially misleading synthetic output publicly (including text on public-interest topics)
See the latest guidance for nuanced exceptions and special cases, including for artistic or editorial content European Commission: Article 50 Q&A.
