Article 50: AI disclosure and content labelling rules, explained

The four duties

1. Providers must inform people when they interact directly with an AI system (unless obvious from context). Your website chatbot must identify itself as AI.

2. Providers of generative systems must mark AI-generated audio, image, video and text output in machine-readable format where technically feasible.

3. Deployers of emotion-recognition or biometric-categorisation systems must inform people exposed to them.

4. Deployers must visibly label deepfakes, and AI-generated text on matters of public interest published without human editorial review.

The December 2026 transition: narrower than you think

The Digital Omnibus gave providers of generative systems already on the market before 2 August 2026 until 2 December 2026 to implement machine-readable marking. That's the only transition. It does not delay chatbot disclosure, deepfake labels, or anything for newly launched systems.

Content generated before 2 August 2026 does not need retroactive labelling.

Quick compliance moves

Rename your chat widget to 'AI assistant' and disclose in the first message. Turn on content credentials (C2PA) where your tools support them. Add visible AI labels to unreviewed published AI text. Document all of it.