EU AI Act & Restaurant Chatbots: Do You Have to Disclose AI in 2026?

The EU AI Act Article 50 requires you to tell customers when they are chatting with an AI bot, not a human. Here is what it means for restaurants and the 2 August 2026 deadline.

Direct Dine team 6 min read AI-assisted

EU AI Act Article 50 is the transparency rule that requires businesses to clearly inform people when they are interacting with an AI system — such as a chatbot — rather than a human, unless it is already obvious. This is not legal advice.

If your restaurant runs an AI chatbot or voice assistant that talks to customers in the EU, this disclosure obligation likely applies.

What does Article 50 actually require?

For AI that interacts with people, you must disclose, at the start of the interaction, that the customer is talking to an AI — clearly and in plain language. A single visible line such as "You are chatting with an AI assistant" generally satisfies the spirit of the rule. AI-generated content can also carry marking obligations.

When is the deadline?

The transparency obligations in Article 50 apply from 2 August 2026. That is the date to have AI-interaction disclosures live on any customer-facing bot. Other parts of the AI Act phase in on different dates, but for restaurant chatbots this transparency date is the one to plan around.

Does it apply to my restaurant chatbot?

If the bot talks to people in the EU — taking orders, answering menu questions, booking tables — and a reasonable customer might think it is human, you should disclose. If it is obviously a machine (a rigid button-only menu with no conversational pretense), the obligation is lighter, but a clear label is still good practice.

What does compliant disclosure look like?

  • A visible "This is an AI assistant" line when the chat opens.
  • Plain language, no buried legalese.
  • Applied consistently across web chat, WhatsApp, and voice.

How Direct Dine handles it

Direct Dine builds AI disclosure into its chat surfaces — the marketplace and business chat, the WhatsApp bot, and AI-generated blog content all carry the Article 50 disclosure helper, so the message is shown automatically before the customer engages. This is not legal advice.

When is this NOT a concern?

  • You run no AI that converses with customers — only human staff reply.
  • Your bot is plainly mechanical and never implies it is a person, though a label is still wise.

If any customer-facing AI is in the loop, treat 2 August 2026 as your disclosure deadline and ship the label well before it.

Keep reading

We value your privacy

We use analytics cookies to understand how visitors interact with our website. No personal data is collected. You can read our Privacy Policy for details.