One more AI Act deadline falls before 2 August: 27 July, 18:00

This analysis first appeared as the weekly Praxikon newsletter.

<h2>The deadline that falls before 2 August</h2> <p>Attention is on 2 August, but there is an earlier date that many organisations are missing. On 10 June the European Commission published the <strong>Code of Practice on transparency of AI-generated content</strong>. To appear on the list of initial signatories, the signatory form must be submitted before <strong>27 July 2026, 18:00 CEST</strong>. That list will be published before 2 August. Signing later remains possible, but you will not be on it.</p> <p>The Code is voluntary. The benefit is evidentiary: following a positive adequacy assessment by the Commission and the AI Board, signatories may rely on the Code to demonstrate compliance with their obligations under Article 50(2), (3) and (5). Those who do not sign remain equally responsible, but must demonstrate compliance by other means and can expect more frequent information requests from supervisory authorities.</p> <h2>The common misreading: Article 50 is not only for builders</h2> <p>One assumption keeps coming back in conversations these past weeks: "we do not build AI, so Article 50 is not ours." That is not correct. The Code deliberately has two sections, and you can sign the second one on its own:</p> <ul> <li><strong>Section 1, providers.</strong> Those developing generative AI systems that produce synthetic text, image, audio or video. Model providers and suppliers of marking and detection technology can also sign this section separately.</li> <li><strong>Section 2, deployers.</strong> Those using generative AI professionally who are affected by the obligations on deepfakes and on AI-generated or manipulated text published to inform the public. This is the category most organisations fall into without realising it: marketing, communications, customer contact, public information.</li> </ul> <p>So the question is not whether you build AI, but which role you hold per use case. That differs from system to system inside the same organisation.</p> <h2>Where the delay stands</h2> <p>The Digital Omnibus was signed on <strong>8 July 2026</strong>, after Parliament (16 June) and the Council (29 June). It was published in the Official Journal on 24 July 2026 (OJ L, 2026/1744) and entered into force on 27 July 2026, the third day after publication. In practice the delay is settled: standalone high-risk Annex III systems move to 2 December 2027, systems embedded in regulated products to 2 August 2028.</p> <p>What does not move stays on 2 August 2026: Article 50, and the enforcement and penalty powers over providers of general-purpose AI models, up to 3% of global turnover or 15 million euro. For marking content already on the market, a transition period runs to 2 December 2026.</p> <h2>Action for this week</h2> <p>Take your list of AI use cases and add one column: are you the provider or the deployer for this one. Put a person's name behind it, not a department. If you use generative AI professionally, weigh this week whether you sign Section 2 of the Code, because that decision has a date: 27 July, 18:00 CEST. The form and instructions are on the European Commission website.</p> <p>If you want that role split and the matching disclosure per use case settled before 2 August, have a look at the <a href="https://embedai.nl/en/diensten/artikel-50-transparantie-check?utm_source=praxikon&utm_medium=referral&utm_campaign=praxikon_weekly&utm_content=editie-2026-07-21">Article 50 transparency check by Embed AI</a>.</p> <p>You receive this weekly AI Act update because you signed up with Praxikon or used one of our tools or templates. You can unsubscribe with one click at the bottom.</p> <p>Until next Tuesday.<br/>Zahed Ashkara<br/>Praxikon</p>