EU AI Act Transparency Obligations: Article 13 & Article 50

Article 13: Transparency to Deployers

Providers of high-risk AI must furnish deployers with sufficient information to use the system correctly and exercise meaningful oversight.

• Provider identity and contact details
• Intended purpose, accuracy, robustness, and cybersecurity properties
• Known limitations and known foreseeable misuses
• Performance with respect to specific persons or groups
• Input data requirements and expected output interpretation
• Human oversight measures and how to exercise them
• Maintenance and lifecycle expectations

Article 50: Transparency to Users & the Public

Article 50 applies more broadly than just high-risk systems. Four core duties:

• AI interaction disclosure — natural persons must be informed they are interacting with an AI system (e.g., chatbots), unless obvious from context
• Synthetic content marking — AI-generated audio, image, video, or text must be machine-readable and detectable as artificial
• Deepfake disclosure — content depicting real persons or events must be disclosed as artificially generated, with limited exceptions for clearly artistic, satirical, or fictional works
• Public-interest text — AI-generated text published to inform the public on matters of public interest must be disclosed unless human review and editorial responsibility apply

How Articles 13 and 50 Interact

The two transparency regimes operate at different levels of the value chain:

Article 13 is a B2B obligation between providers and deployers. Article 50 is end-user-facing. A provider of a high-risk chatbot owes Article 13 documentation to its enterprise deployers AND Article 50 disclosure to the natural persons the chatbot speaks with.