Does Your Website Chatbot Have to Say It’s AI? EU AI Act Article 50 Explained

Yes, in most cases. Since 2 August 2026, the EU AI Act requires that people are told when they are talking to an AI system, such as a website chatbot, an AI agent or an AI phone receptionist, unless it is already obvious. The notice has to be clear and given from the start of the conversation. This is Article 50 of the AI Act, and it applies to systems used in the EU even when the business behind them is based elsewhere. Here is what it means for a business website, based on the European Commission’s own guidance.

What does Article 50 of the AI Act require?

Article 50 is the AI Act’s transparency rule. According to the Commission’s quick facts on the transparency rules, it covers four situations:

SituationWho must actWhat they must do
An AI system talks directly with people (such as chatbots, AI agents and avatars)The provider that builds itDesign it so people are told they are interacting with AI
An AI system generates audio, images, video or textThe providerAdd a machine-readable mark so the content can be detected as AI-generated
Emotion recognition or biometric categorisation is used on peopleThe deployer that uses itInform the people exposed to it
Deepfakes, or AI text published to inform the public on matters of public interestThe deployerDisclose that the content is AI-generated or manipulated

“Provider” means the company that develops the AI system and puts it on the market. “Deployer” means the business that uses it, for example a shop that adds a chatbot to its website.

When do the transparency rules apply?

From 2 August 2026. The Commission’s Article 50 FAQ says the rules cover every system in scope from that date, including ones launched earlier. There is one grace period: generative AI systems already on the market before 2 August 2026 have until 2 December 2026 to add machine-readable marking to their output. That extension came from the Digital Omnibus on AI and covers marking only, not the chatbot disclosure.

For context, the AI Act’s rules for high-risk AI come later: 2 December 2027 for high-risk uses listed in Annex III and 2 August 2028 for AI built into regulated products.

Does a customer service chatbot have to say it’s AI?

Yes. The Commission says people must be notified from the start of the first interaction, in a clear and distinguishable way, when they interact with systems “such as chatbots, AI agents, and avatars”. AI that works only in the background, with no direct conversation with a person, isn’t covered by this duty.

The legal duty sits with the provider, which must design the chatbot so the disclosure happens. In practice that matters to you either way. If you use a ready-made chatbot tool, check that its AI notice is switched on and visible on your site. If a developer builds a custom chatbot for you, the notice should be part of the build.

When is it “obvious” that you’re talking to AI?

No notice is needed when it is obvious that the person is dealing with AI. The Commission says to judge this from the point of view of a reasonably well-informed, observant and circumspect person, and that the exception should be read restrictively. A chat window labelled only “Support” or given a human name and photo is unlikely to count as obvious. When in doubt, add the notice: it is one line of text.

How should a website chatbot tell people it’s AI?

The Act does not prescribe exact wording, and the Commission’s guidelines on Article 50 (adopted July 2026) leave providers and deployers room to choose suitable measures. A practical set-up for a business website:

  1. Say it in the first message, before the visitor types anything: for example, “Hi, I’m an AI assistant for [business]. I can answer questions about orders and opening hours.”
  2. Label the chat window as AI, not only with a name and avatar.
  3. For phone receptionists, say it in the greeting: “You’re speaking with an automated AI assistant.”
  4. Keep the notice visible on mobile and in every language the chatbot speaks.
  5. Offer a route to a person (email, phone or a hand-over). The Act doesn’t require this for chatbots, but it avoids complaints from customers who want a human.
  6. Keep a note of how and where the disclosure appears, with a screenshot, in case anyone asks.

Do AI-written blog posts need a label?

Only in a narrow case. Deployers must label AI-generated or manipulated text that is published to inform the public on matters of public interest, unless it has gone through human review or editorial control and a person or company holds editorial responsibility for it. The Commission notes that a spell-check or grammar pass does not count as human review. For a normal business blog that a person writes or properly edits and publishes under the company’s name, this label duty generally doesn’t apply. Deepfakes are different: realistic AI images, audio or video that show existing people, places or events and could pass as genuine must be disclosed at the latest when someone first sees them.

Does the AI Act apply to businesses outside the EU?

It can. The Commission’s FAQ says providers outside the EU are covered if the output of their AI system is used in the EU. If your website serves customers in Germany, Poland, Lithuania or anywhere else in the EU, treat the chatbot disclosure as required. It is one line of text, so the simplest approach is to show it to every visitor.

What are the fines?

Breaching the transparency obligations can lead to fines of up to €15 million or 3% of total worldwide annual turnover, according to the Commission’s quick facts, which also say the fines are proportionate for SMEs. National authorities in each EU country enforce the rules.

Is there a code of practice?

Yes. A Code of Practice on transparency of AI-generated content was finalised on 10 June 2026. It mainly helps providers with marking and labelling AI-generated content. Signing it is voluntary, but the Commission stresses that the Article 50 duties themselves are legal obligations.

Checklist: is your chatbot ready?

  • The first message says it’s an AI assistant.
  • The chat window or phone greeting is labelled as AI, not only with a human name.
  • The notice shows on mobile and in every language you support.
  • Any AI images or videos of real people on your site are labelled.
  • Customers can reach a person when they need one.
  • You have a screenshot of the disclosure on file.

Frequently asked questions

Does my website chatbot have to say it’s AI?

Yes, in most cases. Since 2 August 2026, the EU AI Act requires that people are told from the start of the conversation that they are interacting with AI, unless that is already obvious to a reasonably well-informed person.

When did the EU AI Act transparency rules start?

Article 50 applies from 2 August 2026. Generative AI systems already on the market before that date have until 2 December 2026 to add machine-readable marking to their output; the chatbot disclosure has no extra grace period.

Does the AI Act apply to UK or US businesses?

It can. The Commission says providers outside the EU are covered if the output of their AI system is used in the EU. If your chatbot talks to people in the EU, add the disclosure.

Do I need to label AI-generated blog posts?

Generally not, if a person reviews or edits the text and your business takes editorial responsibility for it. The label duty covers AI text published to inform the public on matters of public interest without that human review.

Need an AI chatbot that’s compliant from day one?

We build AI chatbots, AI receptionists and AI integrations for business websites, with the disclosure, data handling and hand-over to a person planned in from the start. See our AI solutions service, read how we connect Claude to WordPress with the MCP Adapter, or check an AI-built app with our Lovable app security checklist.

Sources: European Commission: Quick facts, transparency rules for AI systems; European Commission: Transparency obligations under Article 50 (FAQ); Commission guidelines on transparency obligations (20 July 2026); Code of Practice on transparency of AI-generated content; AI Act Service Desk: implementation timeline. Checked 11 November 2026. This is a summary for business owners, not legal advice.

Need help with your website or app?

We build and look after WordPress and WooCommerce websites, mobile apps and AI automations. Tell us what you need and get a free consultation.

Let's discuss your project

We are committed to understanding your requirements and crafting a tailored solution that aligns with your goals.

Enter your details and someone from our team will reach out to find a time to connect with you.

Talk to a Specialist