AI Act: transparency obligations under Article 50 now apply

Chatbots must disclose that they are AI; deepfakes must be labelled. Applies from 2 August 2026.

AWhat changed

From 2 August 2026 the transparency obligations under Article 50 of the AI Act apply: people must know they are interacting with an AI system, deployers must disclose that content is a deepfake, and providers of generative systems must mark outputs in a machine-readable way.

BWhen

From 2 August 2026. For machine-readable marking of outputs from systems placed on the market before 2 August 2026, there is a transition period until 2 December 2026.

CWho it affects

Companies that use a chatbot on their website or in customer support.
Companies that publish AI-generated images, video or audio depicting real people or events.
Providers of generative AI tools.

DWhat to do

Add a clear notice to your chatbot, at the latest at the first interaction, that the user is communicating with AI.
Introduce a rule for labelling deepfake content in marketing and on social media.

Explanation

Article 50 is not part of the postponement for high-risk systems. Exceptions exist, for example, for evidently artistic or satirical works (with an appropriate disclosure) and for AI-generated text that has undergone human editorial review.

Sources

Last checked: 12 October 2026 · How we correct errors