How we use artificial intelligence on this site
This site has an assistant that answers questions, and some of our content is written with the help of language models. This page explains where, in what role, and who is answerable for it — without legal boilerplate where it isn’t needed.
The assistant is an artificial intelligence system
When you write to the assistant on this site, you are not talking to a person. It is an artificial intelligence system: it reads your question, looks for the answer in this site’s documents and pages, and composes a reply using a language model.
What that means in practice:
- It can be wrong. It may misread a question, or answer confidently from outdated information. If a reply looks odd, don’t rely on it — ask a person to confirm.
- It is not medical, tax or legal advice. No answer from the assistant replaces a professional opinion.
- It does not decide for you and does not complete purchases, orders or contracts on your behalf.
- You can always talk to a person. Just ask the assistant, or use the contact details at the bottom of this site.
We are answerable for the content
Whatever tool was used to produce it, every piece of content published on this site is reviewed by a person before it goes online. If a piece of content contains a mistake, the mistake is ours: it is not the AI’s, and it is not an excuse.
Where we do not use it
- We do not generate reviews, testimonials or comments from people who do not exist.
- We do not publish images showing real people in situations that never happened.
- We do not use emotion recognition, biometric categorisation or visitor profiling.
- No automation publishes on its own: the last step is always a person’s approval.
The legal reference
Regulation (EU) 2024/1689, known as the AI Act, sets out transparency obligations in Article 50. The one that applies to this site’s assistant is paragraph 1: whoever makes available an AI system intended to interact directly with people must ensure those people are informed that they are interacting with an artificial intelligence.
Paragraph 4 covers published text instead, and provides an exemption where the content has undergone human review and someone takes editorial responsibility for it. That is the situation on this site, and we state it here anyway.
Article 50 applies from 2 August 2026. Text of the article: artificialintelligenceact.eu/article/50
Found a mistake?
If an answer from the assistant, or any content on this site, contains wrong information, tell us and we will correct it. Reports like that are the most useful feedback we get.