Regulation · AI · Transparency
AI Act Article 50: what needs doing by 2 December 2026
Most of Article 50’s transparency duties have applied since 2 August 2026. For the machine-readable marking of generative AI already on the market, a transition period runs until 2 December 2026. A checklist for developers and operators.
What Article 50 requires
| Duty | Who | In short |
|---|---|---|
| Chatbot disclosure (Art. 50(1)) | Providers | People must be told they are interacting with AI – unless it is obvious |
| Machine-readable marking (para. 2) | Providers of generative AI | Synthetic audio, image, video and text outputs must be marked as artificial, where technically feasible |
| Deep fakes (para. 4) | Deployers | Disclosure; do not rely on the provider’s marking |
| AI text on public-interest matters (para. 4) | Deployers | Disclosure, unless a person with editorial responsibility has reviewed the content |
Disclosures must be clear, appear no later than the first interaction and be accessible. For breaches the source names a ceiling of up to €15 million or 3% of worldwide annual turnover.
The deadlines at a glance
- 2 August 2026: most transparency duties apply.
- 2 December 2026: transition period for machine-readable marking (para. 2) for generative AI systems already on the market before 2 August 2026 – introduced by the “Digital Omnibus” package. Background in The world in 10–20 years.
- 20 July 2026: the Commission published the final guidelines on Article 50.
- June 2026: the voluntary Code of Practice on transparency of AI-generated content was finalised; by the end of July about 190 organisations had signed, according to the source.
Checklist
- Do you have a chat or assistant? Then label it clearly as AI. The chat on this website carries the label “AI Assistant”; the guidelines read the “obvious” exception strictly, so an explicit label is the safe choice.
- Do you generate images, audio, video or text with AI for others? Check whether your provider supplies the machine-readable marking and plan for 2 December 2026.
- Do you publish AI text on matters of public interest? Either disclose it, or have a person with editorial responsibility review the substance. According to the source, the guidelines demand a genuine, knowledgeable review, not a cursory glance.
- Document the process: who reviews, by which criteria, and where is it recorded? My approach is in AI content with human-in-the-loop and the PflegeOasis case.
- Look at the Code of Practice: it is voluntary, but the Commission and the AI Board judged it adequate for paragraphs 2, 4 and 5.
Sources and limits
- Reed Smith: Always read the label – EU AI transparency rules (law-firm article, as of October 2026)
- Usercentrics: Digital Omnibus, high-risk deadlines and Article 50 (secondary source; please check the Official Journal yourself)
Not legal advice. I am a developer, not a lawyer. The details come from the sources listed (as of October 2026); the legal text is what counts.
Keep reading
AI · Labour market · DevOps
The world in 10–20 years: three scenarios instead of a forecast
Nobody knows what the world looks like after AI, robots and drones. What one can do: collect measured numbers, derive scenarios from them and ask what stays the same in all of them – from the view of someone working in IT and DevOps.
AI · LLM · Content pipeline
AI content with human-in-the-loop: 200+ guide articles from LLM drafts
With care law, AI mistakes are not acceptable. This is how I use a language model as a writing aid without handing it the responsibility.
Case study · Health · Guides
PflegeOasis: care-level calculator, objection-letter generator and AI content with human review
A free portal that helps relatives find their way through the German care system – with calculators, checklists and more than 200 guide articles.
Regulation · Security · DevOps
CRA reporting duty from 11 September 2026: 24 hours, 72 hours, 14 days
Since 11 September 2026, manufacturers of products with digital elements must report actively exploited vulnerabilities – in three stages and on a fixed clock. What this means for development and DevOps teams.
More from practice
Find more notes, case studies and projects on the homepage.
All notes Case study MeisterMeld Case study PflegeOasis Case study Kernbohr-Expert Back to the homepage
Comments
Join the discussion: comments run on GitHub Discussions (sign in with a GitHub account). Content from GitHub/giscus is only loaded after you click. Privacy