Free tool · EU AI Act · duties apply 2 August 2026
The Article 50 Duty Mapper
From 2 August 2026, the EU AI Act’s transparency duties go live: chatbots must say they are AI, generated content must be marked, and people exposed to emotion recognition must be told.
Who owes which duty depends on which seat you occupy for each system, and that is the part most coverage gets wrong. Answer a few questions and get your duty map: your seat, your duties, your dates.
Which of these does your organisation run or use?
Pick everything that applies. Each one carries its own duties.
None of these apply? Then Article 50 is probably not your binding constraint. The Annex III high-risk obligations (December 2027) and the rules for general-purpose AI models (like GPT) run on their own tracks. Start with the SME guide.
Why the seat decides everything
Article 50 splits its duties between providers (whoever develops the system, or has it developed, and places it on the market or puts it into service under their own name) and deployers (whoever uses it in a professional context). Interaction disclosure and machine-readable marking fall to providers. Emotion-recognition disclosure and content labelling fall to deployers. A business running a vendor chatbot is typically the deployer, and if it customises or white-labels the tool, it can pick up provider duties too.
The full analysis, including the exemptions and a July checklist, is in the Article 50 guide. For the wider timeline, see EU AI Act compliance for SMEs.
Questions
Who has to comply with EU AI Act Article 50?
Providers and deployers of certain AI systems whose output or users reach people in the EU: AI that interacts directly with people (chatbots, voice agents), generative AI, emotion-recognition and biometric-categorisation systems, and publishers of AI-generated content. The duties apply from 2 August 2026, and the Act reaches non-EU businesses the same extraterritorial way GDPR does.
What is the difference between a provider and a deployer?
The provider develops the AI system, or has it developed, and places it on the market under its own name. The deployer uses the system in a professional context. The split matters because Article 50 assigns different duties to each seat: interaction disclosure and output marking fall on providers, while emotion-recognition disclosure and content labelling fall on deployers. A business that customises or white-labels a vendor tool under its own brand can find itself holding provider duties.
Do the EU’s official AI labelling icons make me compliant?
No. The Commission’s free EU icons (published June 2026) are an optional, user-tested way to present a disclosure. The labelling duty itself is mandatory, you may use the icons or your own label, and using an icon does not by itself discharge the Article 50 duty. The deployer stays responsible for making sure the disclosure genuinely reaches the audience.
My AI system was live before 2 August 2026. Do I get more time?
Only for one duty. Under the Digital Omnibus on AI, Regulation (EU) 2026/1744, in the Official Journal on 24 July 2026 and in force since 27 July 2026, generative systems already on the market before 2 August 2026 have until 2 December 2026 to comply with the machine-readable marking duty in Article 50(2). Nothing else is deferred: the chatbot disclosure, emotion-recognition disclosure, and content-labelling duties all apply from 2 August 2026.
What are the penalties for breaching Article 50?
Up to €15 million or 3% of total worldwide annual turnover, whichever is higher. Prohibited practices, such as workplace emotion recognition (banned since 2 February 2025), carry the top band: up to €35 million or 7% of worldwide turnover.
Built by Michael K. Onyekwere, CIPP/E. Current as at 29 July 2026, reflecting the Digital Omnibus on AI, Regulation (EU) 2026/1744, in force since 27 July 2026. Educational, not legal advice.