Knowledge base

3 min read

AI Act Article 50: transparency for an AI receptionist

Since 2 August 2026, Article 50 of the AI Act requires an AI receptionist to tell callers they are talking to AI. Who carries that duty, and what you check.

  • AI Act
  • transparency
  • regulation
  • AI receptionist

Since 2 August 2026, the European AI Regulation (Regulation (EU) 2024/1689) applies to most AI applications. One of the obligations that takes effect is in Article 50: transparency. Anyone deploying an AI receptionist must let the people calling know they are talking to an AI.

What is the AI Act?

The AI Act is a risk-based framework from the European Union. Not every AI application gets the same rules; the higher the risk, the heavier the obligations:

  • Prohibited practices (Article 5) — such as social scoring. Not allowed.
  • High risk (Annex III) — for example AI that makes decisions about people in hiring or lending. Heavily regulated.
  • Limited risk — AI that communicates with people. The transparency obligation of Article 50 applies here.
  • Minimal risk — for example spam filters. No specific obligations.

An AI receptionist that answers calls, schedules appointments and routes callers falls under limited or minimal risk in practice. That means: no heavy licensing requirement, but the transparency obligation does apply.

What exactly does Article 50 say?

Article 50(1) puts the duty on the provider: whoever builds an AI system that interacts directly with people must design it so that those people are told they are interacting with an AI, unless this is obvious from the circumstances. For a chatbot or AI receptionist, that is the supplier of the system. If your business deploys such a system, you check that it actually gives that notice. As a deployer (the business that uses the system), you only have duties of your own for emotion recognition or biometric categorisation (paragraph 3) and for deepfakes or text on matters of public interest (paragraph 4).

For an AI receptionist this means, concretely: at the start of the call, the caller must hear that they are speaking with an AI receptionist. In a phone call this is not "obvious" — you cannot see who is on the other end — so the disclosure is required.

The caller does not need to be cut off; it is about a clear, honest disclosure upfront. The receptionist can then simply schedule the appointment or route the call.

When does this apply?

The AI Regulation entered into force on 1 August 2024, with a phased application. The transparency obligations of Article 50 fall under the general application and have applied since 2 August 2026. Anyone setting up an AI receptionist today is therefore doing so in a way that already complies with Article 50.

Relationship to the GDPR

The GDPR continues to apply as before. The AI Act comes on top of it; it does not replace it. For an AI receptionist this also means following the GDPR principles: process only what is necessary, and do not store data longer than needed. Separate rules apply especially when phone calls are recorded or transcribed.

Please note

This article is a general explanation, not legal advice. The AI Act is new and its practical application is still being shaped. For specific questions, consult a lawyer specialised in privacy and AI regulation.

Let's build something.

Got an idea, a problem or a project? I'm happy to help. No obligations, just a good conversation.

  • No obligation
  • 30 minutes
  • No sales talk
  • Actionable advice