Article by Article Part 3 of 3
  1. 1 Reading the AI Act One Article at a Time
  2. 2 The EU Has an AI Act, the UK Has Regulators and No Bill
  3. 3 Article 50 of the EU AI Act Makes AI Declare Itself
About this series

Article 50 of the EU AI Act Makes AI Declare Itself

Article 50 of the EU AI Act Makes AI Declare Itself
Generated with AI, with its C2PA provenance mark intact.

Article 50 is the part of the EU AI Act that most software teams meet first. It has applied across the European Union since 2 August 2026, and the one grace period it got, for marking the output of generative systems already on the market by then, ends on 2 December. It reaches beyond the EU's borders too: a provider based anywhere is covered once its system is on the EU market or its output is used in the Union, so a UK or US product with users in the EU has to comply (the note on territorial scope has the limits). In the UK itself it is not law, and the UK has regulators and no AI bill. Article 50 is also short enough to read in full, so that is the format: the article verbatim below, every clause of it selectable, and the notes in the panel beside it.

Each clause has a plain reading of what it means. On top of that sit notes of five kinds, each with its own colour: what it changes in software, what is contested, what the Commission's guidance says, where it connects to the rest of the law, and dates. On a phone, tap a clause and the notes slide up from the bottom. With a keyboard, Tab moves through the clauses, and once one is open, j and k step forwards and back. Every clause and every note has its own link, so you can send someone straight to the sentence you are arguing about.

Two documents from the summer do most of the interpreting. The Commission's guidelines on Article 50 came out on 20 July 2026, and the Code of Practice on Transparency of AI-Generated Content was finalised on 10 June and has since been assessed as adequate by the Commission and the AI Board. The notes cite both by paragraph or measure number, so every claim can be checked against the source.

If you only open three clauses, make them these. The design duty in paragraph 1, because a line in your terms and conditions does not meet it. The marking duty in paragraph 2, because if your product calls a model API and shows the result under your name, the duty is yours. And the standard editing exception, because whether an output needs a mark depends on which feature produced it, so the decision lives in your code.

The notes keep growing after today. Each one is dated, and the log at the foot of the page lists what was added when. The compliance-as-code post from February now carries a dated note pointing here, since its middleware numbered the paragraphs differently from the way the guidelines read them.

This is an engineer's reading and no substitute for legal advice. Lawyers I know are reviewing it, and their corrections will appear in the notes with their dates.

This page's own Article 50 notices

A page explaining Article 50 is covered by Article 50, so here are its notices. You rarely see a box like this anywhere, partly because enforcement has barely started and partly because nobody enjoys writing one. Expect to see more of them once the regulators get going. It reads a little clunky, which is roughly what compliance looks like when it is done in public.

  • The text was written with AI assistance. An article about the law, published to inform the public, is the kind of text paragraph 4 covers. It relies on the editorial exception: I review everything before it is published, fact-checking included, and I hold editorial responsibility for it. I am Andy Cross, and the About page says who I am.
  • The notes were edited with AI after they went live. On 20 November an AI pass over the wording changed ten clause readings and nine notes, and corrected three notes that went further than the sources allow (each says what changed). The guidelines say a substantive AI change after sign-off voids the editorial exception, so every change was reviewed and signed off again before it was republished. The note on sign-off describes exactly this.
  • The cover picture is AI-generated. There is no such chambers off Chancery Lane. The file still carries the generator's machine-readable C2PA mark after upload, which the note on metadata explains is not a given.
  • Whether a personal blog is a deployer at all is arguable, since Article 3(4) leaves out personal, non-professional use. The notices are here anyway. They cost a few lines.

EU AI Act · Chapter IV

Article 50

Transparency obligations for providers and deployers of certain AI systems

Providers shall ensure that AI systems intended to interact directly with natural persons are designed and developed in such a way that the natural persons concerned are informed that they are interacting with an AI system, unless this is obvious from the point of view of a natural person who is reasonably well-informed, observant and circumspect, taking into account the circumstances and the context of use. This obligation shall not apply to AI systems authorised by law to detect, prevent, investigate or prosecute criminal offences, subject to appropriate safeguards for the rights and freedoms of third parties, unless those systems are available for the public to report a criminal offence.

Providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content, shall ensure that the outputs of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated. Providers shall ensure their technical solutions are effective, interoperable, robust and reliable as far as this is technically feasible, taking into account the specificities and limitations of various types of content, the costs of implementation and the generally acknowledged state of the art, as may be reflected in relevant technical standards. This obligation shall not apply to the extent the AI systems perform an assistive function for standard editing or do not substantially alter the input data provided by the deployer or the semantics thereof, or where authorised by law to detect, prevent, investigate or prosecute criminal offences.

Deployers of an emotion recognition system or a biometric categorisation system shall inform the natural persons exposed thereto of the operation of the system, and shall process the personal data in accordance with Regulations (EU) 2016/679 and (EU) 2018/1725 and Directive (EU) 2016/680, as applicable. This obligation shall not apply to AI systems used for biometric categorisation and emotion recognition, which are permitted by law to detect, prevent or investigate criminal offences, subject to appropriate safeguards for the rights and freedoms of third parties, and in accordance with Union law.

Deployers of an AI system that generates or manipulates image, audio or video content constituting a deep fake, shall disclose that the content has been artificially generated or manipulated. This obligation shall not apply where the use is authorised by law to detect, prevent, investigate or prosecute criminal offence. Where the content forms part of an evidently artistic, creative, satirical, fictional or analogous work or programme, the transparency obligations set out in this paragraph are limited to disclosure of the existence of such generated or manipulated content in an appropriate manner that does not hamper the display or enjoyment of the work.

Deployers of an AI system that generates or manipulates text which is published with the purpose of informing the public on matters of public interest shall disclose that the text has been artificially generated or manipulated. This obligation shall not apply where the use is authorised by law to detect, prevent, investigate or prosecute criminal offences or where the AI-generated content has undergone a process of human review or editorial control and where a natural or legal person holds editorial responsibility for the publication of the content.

The information referred to in paragraphs 1 to 4 shall be provided to the natural persons concerned in a clear and distinguishable manner at the latest at the time of the first interaction or exposure. The information shall conform to the applicable accessibility requirements.

Paragraphs 1 to 4 shall not affect the requirements and obligations set out in Chapter III, and shall be without prejudice to other transparency obligations laid down in Union or national law for deployers of AI systems.

The Commission shall encourage and facilitate the drawing up of codes of practice at Union level to facilitate the effective implementation of the obligations regarding the detection, marking and labelling of artificially generated or manipulated content. The Commission, taking utmost account of the opinion of the Board, shall assess whether adherence to those codes of practice is adequate to ensure compliance with the obligations laid down in paragraphs 2 and 4 of this Article, in accordance with the procedure laid down in Article 56(6). If it deems the code of practice to be inadequate, the Commission may adopt an implementing act specifying common rules for the implementation of those obligations in accordance with the examination procedure laid down in Article 98(2).

Text from Regulation (EU) 2024/1689, the AI Act, consolidated with the Omnibus, checked 1 November 2026. Engineer's reading, not yet reviewed by a lawyer.

Changes to these notes

The notes grow as things come up, and each one carries the date it was added. Last addition 1 November 2026.

1 November 2026, 38 notes

Sources