Article 50 · Chapter IV: Transparency Obligations for Providers and Deployers of Certain AI Systems
Transparency Obligations for Providers and Deployers of Certain AI Systems
Summary
Cross-cutting transparency duties for systems regardless of risk tier. (1) Providers of AI systems intended to interact with natural persons must inform users they are interacting with AI unless obvious from context. (2) Providers of generative AI must mark synthetic audio/image/video/text content as artificially generated in machine-readable format. (3) Deployers of emotion recognition or biometric categorisation systems must inform exposed persons. (4) Deployers of deep fakes must disclose; for text published to inform the public on matters of public interest, the AI-generated nature must be disclosed unless human-reviewed and an editorial responsibility is taken.
Key obligations
- Disclose AI interaction to natural persons (chatbots, voice agents)
- Watermark / mark generative AI outputs in machine-readable form
- Inform exposed persons of emotion recognition or biometric categorisation
- Disclose deep fakes; disclose AI-generated public-interest text