Article 50 AI disclosure generator
A starting point and a self-assessment, not legal advice. Adapt the wording to your product and your users. This page cannot see your product, so it cannot tell you whether your notice is enough.
Live preview
The notice in a mock of your interface, before the first interaction. It updates as you type.
Disclosure sentence
HTML snippet
Copy-paste HTML with a few lines of CSS and no dependencies. role="note", the word AI in the text, readable on a phone, light or dark from the visitor's setting. Place it where people see it before or at the first interaction.
Privacy policy or terms clause
A draft. Fill in the parts in square brackets. It supports the notice, it does not replace it.
What Article 50 asks for this kind of system
Who owes what: the provider builds or has the AI system built and puts it out under its own name; the deployer uses it in its own activity. A startup shipping its own chatbot on a vendor's model is usually the provider of that chatbot. Tick what you have done; nothing is saved.
Article 50 sits beside the risk tiers, not instead of them. If the same system is used for an Annex III purpose, check the high-risk checker too.
Frequently asked questions
Do I need an AI notice if it is obvious that users are talking to a bot?
Article 50(1) drops the notice only where it is obvious to a reasonably well-informed, observant and circumspect person. Read that narrowly. A chat widget with a human name and photo, or a voice that sounds like a person, is designed not to be obvious. If in doubt, show the notice: it costs one line.
Where should the notice appear?
Where the interaction starts, at the latest at the first interaction or exposure (Article 50(5)): the chat greeting, the first words of a call, next to the generated image. A paragraph in your privacy policy supports it but does not replace it, because people do not read the policy before they start.
Is a visible "AI-generated" label enough for images, audio and video?
Not for the provider of the generator. From 2 December 2026, Article 50(2) asks providers to mark synthetic output in a machine-readable format so it can be detected as AI-generated, as far as technically feasible. A visible label helps people; marking is technical, such as metadata, a watermark or both.
Who has to show the notice: us or our model vendor?
Article 50(1) puts the interaction notice on the provider of the AI system. If you build a chatbot on a vendor's model and offer it under your own name, you are usually the provider of that chatbot, so the notice is yours to build in. Article 50(4) duties for deepfakes and AI text on matters of public interest sit with the deployer: whoever publishes the content.
Is this template all I need for Article 50?
No. It gives you wording, a component and a checklist to start from. Whether Article 50 applies, which paragraphs, and whether you or your model vendor is the provider depend on your product. Treat the output as a draft and keep a dated note of what you decided and why.
Check the rest of your obligations
This tool answers one question. The free scan takes about two minutes: eight questions plus your email, and your likely risk tier on screen. The gap report then lists the Articles and deadlines that apply to your product.
No card, no account. Gap report from €59. Prices exclude VAT, added at checkout where applicable. 14-day refund.
Self-assessment, not legal advice. This page and the Reglynn report are a structured self-assessment based on the EU AI Act (Regulation 2024/1689, as amended by the June 2026 Omnibus). They indicate where you likely stand and what to verify with a qualified advisor. Reglynn is not a notified body and issues no certification.
How this was made. Reglynn reports are generated by an AI system built on Anthropic's Claude and delivered automatically. Reglynn is a Limited Risk AI system under Article 50; we disclose AI involvement wherever it occurs, including in our chat assistant. Our own self-assessment.