Guide
Does your website need an AI notice? Five checks
If your website has an AI chatbot, writes with AI or sends data to an AI service, visitors should be told. Five practical checks and a two-minute self-check.
The short answer
If your website or app uses AI where visitors can meet it, tell them in plain words. Five checks cover most cases. First, list every place AI is involved: a chatbot, a writing or translation tool, a recommendation, a form that is read by AI. Second, tell people when they are talking to AI and not to a person, at the moment they start. Third, label content that was generated or changed by AI where a reader could mistake it for human work, for example synthetic images or voices. Fourth, know what data leaves your site for an AI service, and say so in your privacy page. Fifth, make sure your team knows what the tools may and may not be used for. Rules such as the EU AI Act ask for this kind of transparency in many cases, and the details depend on your country and use. This guide is a practical check, not legal advice.
In this guide
Who this is for
You added an AI chatbot, let a tool draft your texts or images, or send customer messages to an AI service to sort or answer them. It works. The open question is whether visitors know. Being open about it builds trust, and in a growing number of places the law asks for it.
You do not need technical knowledge for most of these checks. The data check may need whoever built the site.
The five checks
List where AI is involved. Write down every place: a chat window, an assistant that answers emails, an AI tool that writes or translates pages, generated images, a search or recommendation feature, a form whose messages are summarised by AI. Include tools your team uses quietly. You cannot tell anyone about something you have not listed.
Say when someone is talking to AI. If a visitor chats with a bot, or hears a synthetic voice, they should know at the start, in plain words such as “You are chatting with an AI assistant” and a way to reach a person. Put it in the chat window itself, not only in a footer page.
Label AI-made content where it could pass as human. Realistic images, voices, video or testimonials created or changed by AI can mislead. Add a short label next to them, for example “Image created with AI”. Text you wrote with an AI tool and then edited and checked yourself is usually different from text published as it came; decide what you can stand behind and be honest about it.
Know what data leaves your site. When a visitor types into an AI feature, where does the text go, who processes it, and is it kept or used for training? Check the settings and terms of the provider and describe this in your privacy page in simple words. Avoid sending personal data to an AI service you have not reviewed.
Prepare your team. Agree in a short note what staff may put into AI tools (no customer secrets, no personal data in free tools), who checks AI output before it is published, and who answers when a customer asks. The EU AI Act also asks organisations that use AI to make sure their people have enough understanding of it (EU AI Act overview).
A two-minute self-check
- Open your site as a visitor: where does AI appear? Is it announced?
- Start the chatbot: does it say it is an AI and offer a way to reach a person?
- Look at images and videos: could any be mistaken for a photo or a real person?
- Read your privacy page: does it mention the AI service and what is sent to it?
- Ask one colleague what they may paste into AI tools. Do the answers match?
When to get help
Ask someone experienced if:
- you are not sure which AI features your site and tools really use
- personal or sensitive data reaches an AI service
- you serve customers in several countries and are unsure which rules apply
- you want a lawyer’s view on the exact wording, which is always worth it for sensitive cases
We cover the practical side in our AI readiness and disclosure pack and can go through your site with you in a free 30-minute video call: book a call. For the legal wording, ask a lawyer.