EU AI Act: telling people when they deal with AI (Article 50)
BuildAIPolicy Research · Added 27 October 2026 · Last checked 27 October 2026
What it is
People must be told when they interact with an AI system such as a chatbot, unless it's obvious, and AI-generated content such as deepfakes must be disclosed. Providers of generative AI already on the market before 2 August 2026 have until 2 December 2026 to mark its outputs.
Who it affects
Organisations that offer chatbots or publish AI-generated content in the EU, and providers of generative AI.
What to consider
- Add a clear notice to any chatbot or AI assistant that customers use.
- Label AI-generated images, audio or video that could be mistaken for real.
Sources
Put rules like this into practice
The AI Policy & Risk Pack gives your organisation an AI use policy, a risk register, an AI tool inventory and notice wording, written from how you use AI. Preview every section free before you pay. It isn't legal advice.
A general summary of public information, checked against the sources listed on the date shown. It isn't legal advice and may not cover every rule that applies to you. Check the sources, and get professional advice for your situation.