# Do I have to tell customers they are talking to an AI? What Article 50 of the EU AI Act says about telling people they are dealing with an AI, and who the article addresses. Published: 2026-10-12 Canonical: https://felesh.ai/blog/do-i-have-to-tell-customers-they-are-talking-to-an-ai ## The short answer The EU AI Act has a rule on this. It is Article 50(1) of Regulation (EU) 2024/1689, which was published in the Official Journal of the European Union on 12 July 2024. It says providers must ensure that AI systems intended to interact directly with natural persons are designed and developed in such a way that those persons are informed that they are interacting with an AI system. The exception is where this is obvious from the point of view of a reasonably well-informed, observant and circumspect natural person, taking into account the circumstances and the context of use. The article is addressed to providers: broadly, whoever develops an AI system, or has one developed, and puts it on the market or into service under its own name (Article 3(3)). Whether a particular business is a provider depends on the facts; see below. Whether that rule reaches a particular business, and how, depends on the business's role under the Act and on the facts. This post sets out what the text says. It does not say that any set-up, ours included, meets it. This is general information, not legal advice. ## When does it apply, and to whom? Article 113 says the Regulation applies from 2 August 2026, with listed exceptions for other parts of it. The European Commission states that the Article 50 transparency rules apply from 2 August 2026. Article 2(1) sets out the scope. It covers, among others, providers placing AI systems on the market or putting them into service in the Union, deployers that are established or located in the Union, and providers and deployers in a third country where the output of the AI system is used in the Union. Regulation (EU) 2026/1744 has since amended the Act. It changed Article 50(7) and added a transition to 2 December 2026 for the marking duty in Article 50(2). It did not change paragraphs 1 or 5. ## What does the rule say about timing? Article 50(5) says the information referred to in paragraphs 1 to 4 must be provided to the natural persons concerned in a clear and distinguishable manner, at the latest at the time of the first interaction or exposure. It must also conform to the applicable accessibility requirements. ## How is "obvious" tested? Article 50(1) does not ask whether the business finds it obvious. The test it states is the point of view of a natural person who is reasonably well-informed, observant and circumspect, taking into account the circumstances and the context of use. The text frames this as an exception: the information is to be given "unless this is obvious". The circumstances and context of use are what the test looks at. Article 50(1) also has a carve-out for AI systems authorised by law to detect, prevent, investigate or prosecute criminal offences, subject to appropriate safeguards. ## Does the duty fall on the business or on the tool maker? Article 50(1) is addressed to providers. Article 3(3) defines a provider as a natural or legal person, public authority, agency or other body that develops an AI system, or has one developed, and places it on the market or puts it into service under its own name or trademark, whether for payment or free of charge. Article 3(4) defines a deployer as a natural or legal person, public authority, agency or other body using an AI system under its authority, except where the system is used in the course of a personal non-professional activity. Other paragraphs of Article 50 address deployers in particular cases: emotion recognition and biometric categorisation systems (paragraph 3), deep fakes (paragraph 4, first subparagraph), and AI-generated text published to inform the public on matters of public interest (paragraph 4, second subparagraph, with an exception where the text has undergone human review or editorial control and someone holds editorial responsibility). Article 50(6) adds that paragraphs 1 to 4 do not affect the requirements in Chapter III and are without prejudice to other transparency obligations laid down in Union or national law for deployers of AI systems. Which role a given business holds for a given tool depends on how the tool is developed, put on the market or put into service, and used. That is a question for an adviser who has the facts. ## What does this look like in a Felesh inbox? A customer writes from the chat on your website, or from Instagram, WhatsApp or Telegram. The message arrives in one inbox. 1. Felesh answers from what you taught it: the documents you uploaded and the instructions you wrote. 2. If it cannot answer, the message goes to "Needs you", with the reason it could not. 3. Nobody is notified, so you look in "Needs you" yourself. 4. When you join the conversation, Felesh stays quiet. That walk-through contains no statement about what the customer was told at the start of the conversation. Felesh answers from what you give it, so the wording is your wording and your documents. It does not settle the disclosure question. Article 50(5) speaks of information provided at the latest at the time of the first interaction. We make no claim that any set-up meets Article 50. ## Where to read more - Article 50 and Article 3 of [Regulation (EU) 2024/1689](https://eur-lex.europa.eu/eli/reg/2024/1689/oj), in the Official Journal on EUR-Lex, in the language you are most comfortable with. - The [European Commission's guidelines on the transparency obligations](https://digital-strategy.ec.europa.eu/en/library/guidelines-transparency-obligations-providers-and-deployers-ai-systems) under Article 50, published on 20 July 2026, which it says clarify the scope, the relevant legal definitions, the obligations and the exceptions. If you want to see how customer messages, answers and hand-overs are laid out, read about the shared inbox. This is general information, not legal advice.