Home / EU AI Act Article 50, in plain language
Article 50 applies from 2 August 2026

EU AI Act Article 50, in plain language

Article 50 is the transparency chapter of the EU AI Act. It does not ban anything. It says that when people meet AI, they have to know it. This guide breaks down each obligation and the date it starts to bite.

Most of the EU AI Act is about risk classification and paperwork for a small number of high risk systems. Article 50 is different. It touches almost every business that puts an AI feature in front of the public, from a support chatbot to an image generator. If a person can interact with your AI, or see something your AI produced, Article 50 has a rule for it.

The four transparency duties

Article 50 sets out four distinct obligations. They apply to different actors, so read them against what you actually run.

  1. Talking to AI (Art. 50(1)). If people interact with an AI system, you must tell them, unless it is obvious from the context. A chatbot that could pass for a human agent is the classic case.
  2. Synthetic media (Art. 50(2)). Providers of systems that generate audio, image, video, or text must mark the output as artificially generated, in a machine readable way where technically feasible.
  3. Deepfakes (Art. 50(4)). Anyone who deploys a system to create deepfakes must disclose that the content is artificially generated or manipulated.
  4. Public interest text and biometrics. AI generated text published to inform the public on matters of public interest must be flagged, and emotion recognition or biometric categorisation must be disclosed to the people exposed to it.

When does it apply?

The transparency obligations in Article 50 apply from 2 August 2026. That date was not pushed back by the Commission simplification package. The one narrow easing is for the machine readable marking of synthetic output under Art. 50(2): systems already on the market before 2 August 2026 get until 2 December 2026 to add it. The duty to tell a person they are talking to AI, and to label deepfakes, starts on 2 August with no grace period.

Short version: if your chatbot goes live to EU users, the disclosure notice needs to be there on 2 August 2026.

Who has to comply?

Article 50 follows the AI, not your postcode. If your system reaches users in the EU, the obligations apply even if your company sits elsewhere. Both providers (who build the system) and deployers (who put it to use on their own site) carry duties, which is why a shop running a third party chatbot still has to show the notice.

What a compliant notice looks like

The law does not hand you exact wording, but it does set the bar: the disclosure has to be clear, given at the latest at the first interaction, and accessible. In practice that means a short, plain notice, in the user's language, that a real person can perceive, including via screen reader. A line buried in a privacy policy nobody opens does not meet the timing requirement.

Common questions

Is Article 50 delayed?

No. The transparency obligations apply from 2 August 2026. Only the machine readable marking of synthetic content under Article 50(2) has a short grace window to 2 December 2026 for systems already on the market before August.

Does a simple support chatbot count?

Yes, if a user could reasonably think they are talking to a person. That is exactly the situation Article 50(1) was written for.

What if I only serve customers outside the EU?

Then Article 50 does not reach you. The moment EU users can interact with the system, it does.

Show the Article 50 notice on your site

Paste one snippet. It finds your chatbot, shows a clear and localized disclosure, and keeps the wording current as guidance changes.

Start free