Does your chatbot need to say it's AI?
Yes — if your chat widget is powered by AI and a reasonable visitor might think they're talking to a person, you must tell them. Article 50(1) has been enforceable since 2 August 2026. The fix is usually one line of text in the widget's opening message.
What the law requires
Article 50(1) requires providers to design AI systems intended to interact directly with people so that those people are informed they are interacting with an AI system — unless it is obvious from the circumstances to a reasonably well-informed and observant person.
The "obvious from context" exemption is narrower than most site owners assume. A widget branded "AI Assistant" that opens with "Hi, I'm an AI assistant" is clearly covered. A widget labelled only "Chat with us" that produces fluent, human-sounding replies is not — the fluency is precisely what makes the disclosure necessary.
The obligation sits with the provider of the system, but a deployer who puts a third-party chatbot on their site and strips or hides its disclosure takes on exposure of their own. In practice, if the widget is on your domain and says nothing, you are the one a market surveillance authority will contact.
What our scanner looks for
- Known AI chat and voice vendors by their script signatures — Intercom, Drift, Crisp, Tidio, HubSpot Conversations, Zendesk, Tawk.to, LiveChat, Freshchat, Botpress, Voiceflow, Chatbase, Ada, Algolia Ask AI, ElevenLabs Conversational AI, and generic "Ask AI" builds.
- Custom implementations, via chat-widget markup and endpoints such as /api/chat.
- Whether disclosure wording appears near the widget, in the page text, or in your legal pages — read in the page's own language, not just English.
How to fix it
- 1Put the disclosure in the widget's first message, where it is read before anyone types: "You're chatting with an AI assistant. Ask for a human any time."
- 2Name it in the widget header or launcher label too — "AI Assistant" rather than "Chat".
- 3Offer a route to a human and say so. It is not required by Art. 50(1), but it is what turns a legal disclosure into something a customer values.
- 4Add a line to your privacy policy describing the AI chat, what it processes and which provider runs it. That is where a regulator or a careful buyer looks second.
Deadlines and penalties
Enforceable since 2 August 2026. Enforcement is by the market surveillance authority of each Member State where the system is placed on the market or used. Penalties under Art. 99(4) reach €15 million or 3% of total worldwide annual turnover, whichever is higher.
Free, no account needed. We'll show you every page where this applies, with the evidence.
Common questions
Our chatbot is rule-based, not AI. Does it still apply?
Article 50(1) covers AI systems. A deterministic decision-tree bot with no model behind it falls outside it. The distinction is not always clean, though: many "rule-based" widgets now route to an LLM for fallback answers, and that part is in scope.
Is a line in the privacy policy enough?
Almost certainly not on its own. The obligation is to inform the person interacting with the system, at the point of interaction. A disclosure buried three clicks away does not reach someone who opened a chat bubble and started typing.
We're outside the EU. Does this apply to us?
The AI Act applies where the system's output is used in the EU, regardless of where the provider is established. A US company whose chatbot serves EU visitors is in scope.
Related
This page describes technical indicators and the obligations they relate to. It is not legal advice. A website scan detects indicators, not legal compliance — which is why our reports say “no issues detected” and never “compliant”.