An AI chatbot on your website.

Article 50 of the AI Act — the transparency rules — has applied since 2 August 2026. For a website chatbot, the core duty is in Article 50(1): an AI system that interacts directly with people must be designed so that they are informed they are dealing with an AI, unless that is already obvious to a reasonably well-informed person. Article 50(5) adds that the information must be clear, given at the latest at the first interaction, and meet accessibility requirements.

Two other parts can matter. AI-generated text published to inform the public on matters of public interest must be disclosed as such, unless a person has reviewed it and someone holds editorial responsibility. And providers of systems that generate content must mark it in a machine-readable way, with systems already on the market before August given until 2 December 2026.

Breaches of Article 50 can be fined up to 15 million euros or 3% of worldwide annual turnover, whichever is higher — and for small and medium-sized businesses, whichever is lower. This is a summary, not legal advice; if your obligations are unclear, ask a lawyer who practises in this area.

For most chatbots the disclosure is a sentence in the chat window before the first message — “You're chatting with an AI assistant” — and a label that stays visible, written so a screen reader announces it. Avoid human names and photos that suggest a person is typing.

Look at the questions people actually ask. If most are covered by your existing pages, a better FAQ, clearer navigation or a search box may solve the problem for less money and less risk. A chatbot earns its place when questions are numerous, varied and answerable from material you already have.

If you go ahead, test it against a set of real past questions before launch and again whenever the content or model changes, so improvements in one place do not quietly break answers somewhere else.