AI Act Article 50: what you must label now

Chatbots and voice assistants

Providers must design AI systems that interact directly with people so that people learn they are dealing with AI, unless this is obvious to a reasonably well-informed person (Art. 50(1)).

AI-generated images, video, audio and text

Providers of systems generating such content must mark it in a machine-readable format as artificially generated (Art. 50(2)). For new systems this applies since 2 August 2026; for systems already on the market, from 2 December 2026.

Publishing deepfakes and AI texts

Deployers publishing realistic images, videos or voices must disclose that they are artificially generated. For art, satire and fiction a notice that does not hamper the work is enough. AI texts informing the public on matters of public interest must also be disclosed, unless a person reviewed them editorially and holds responsibility (Art. 50(4)).

What to do now

• List all AI systems and AI content in your company
• Add a clear notice to chatbots, for example "You are chatting with an AI assistant"
• For your own generative features, build in technical marking such as metadata or watermarks
• Set a rule for marketing and editorial teams on when AI content is labelled
• Run the check above and document the result