DevalandDevaland

Free self-check, 3 minutes

Does your chatbot or voice assistant say it is an AI?

Ten questions about what your assistant actually says and shows. You get a list of technical observations: what to read, in which channel, and what a recorded test would capture.

This is not legal advice. The tool gives technical observations only. It does not tell you whether you comply with Article 50 of the EU AI Act or any other law, and it does not say what the law requires of you. Your lawyer or the competent authority draws that conclusion. Your answers stay in your browser: nothing is sent to us.

1. The very first thing the assistant says or shows, before the visitor asks anything, states that it is an AI or an automated system. (A label such as "virtual assistant" or "the assistant" alone does not say it.)
2. The same wording is used in every channel the assistant runs on (website chat, voice widget, phone line, WhatsApp or other).
3. The assistant states it is an AI in the same language it then speaks to the visitor, in every language it serves.
4. On a phone line or voice widget, the AI statement is SPOKEN at the start of the call (not only written on a web page).
5. When asked "am I talking to a human?", the assistant says plainly that it is an AI.
6. The assistant does not use a human name, photo or voice in a way that suggests a person is answering.
7. The assistant produces images, audio, video or documents that visitors can download or share.
8. After every change to the assistant (prompt, model, voice), someone re-reads the opening wording in each channel and keeps a dated record of it.
9. The assistant quotes prices, delivery or refund terms, and someone regularly compares what it says with your own published page.

Official sources.

Limits, stated plainly

  • The result is built only from your answers. We do not open your assistant, and we do not verify any answer.
  • It lists what to read and record. It does not say whether anything is allowed, required or sufficient. We do not interpret the AI Act.
  • A "Yes" on every question is not proof that your assistant says what you think it says. Reading it is.
  • We promise no result. This is not a compliance assessment and not legal advice.

More free tools

Opens in a new tab