Yes — in almost all cases. If someone could reasonably think they’re chatting with a person, you have to make clear it’s AI, right there in the conversation, from 2 August 2026.
Article 50 of the EU AI Act requires that people interacting with an AI system are told they’re dealing with AI. The only get-out is where it’s genuinely obvious to a reasonable person — and regulators read that narrowly, so it’s not something to lean on. It covers website chat widgets, WhatsApp bots, and AI voice or phone agents alike.
It must be in the interaction itself, at or before the first message — not buried in your terms and conditions or privacy policy. A vague label like “Assistant” isn’t enough on its own.
A short, plain line does the job: “Hi, I’m an AI assistant. I can help with most things, and I’ll pass you to a human any time.” For voice, an audible opener — “You’re speaking with an AI assistant.” Best practice is a visible “AI” badge plus that first-message notice. You don’t need a big legal disclaimer.
If you use a third-party bot (Intercom, a plugin, and so on), don’t assume the vendor switched the disclosure on. What matters is that your users are actually told — so check it’s showing.
Breaches of Article 50 can draw fines of up to €15 million or 3% of worldwide annual turnover, whichever is higher.
Not sure where you stand? The free checker tells you in two minutes which duties apply to you — no signup, no data collected.
Run the free 2-minute check →Other common questions