Praxikon

Q&A

Transparency obligations under Article 50 of the AI Act - Questions and Answers

PublishedEuropean Commission, DG CONNECT24 Jul 2026

Official questions and answers on the guidelines to Article 50. The page addresses who is provider and who is deployer, when a system must disclose that the user is communicating with AI, what requirements apply to machine-readable marking of synthetic content, and how the discoverability of deepfakes and AI-generated text on matters of public concern must be handled. The exception for human editorial oversight, the way compliance can be demonstrated and enforcement are also covered. The page confirms that a limited transition period applies to systems placed on the market before 2 August 2026, exclusively for the marking and detection obligation of Article 50, paragraph 2, to which they are only required to comply from 2 December 2026. The date 24 July 2026 is the last-updated date.

What this means for you

Article 50 has been in force since August 2026 and affects almost every organisation deploying chatbots or generative AI, so the official delineation of the exceptions is directly relevant.