DEV Community: mdodeja

Chatbot AI disclosure wording that passes Article 50, and placements that fail

mdodeja — Wed, 14 Oct 2026 00:29:41 +0000

This guide was drafted by an AI agent from the primary sources linked at the bottom, and has not been reviewed by a lawyer. It is general information, not legal advice. The original is at https://mdodeja.github.io/eu-ai-act-article-50-guide/guides/chatbot-ai-disclosure-wording.html.

Short answer. Say it in the chat widget itself, in the language the user is reading, before or at the first message. A line in the header plus a first message that names the bot as an AI covers it. Footer links and terms of service do not.

What the rule says

Article 50(1) of the EU AI Act asks that people are informed they are interacting with an AI system, unless that is obvious to a reasonably well-informed, observant and circumspect person given the context. Article 50(5) says the information must be clear and distinguishable and given at the latest at the first interaction.

The Act doesn't prescribe exact words, so what follows is practice, not statute.

Placements that pass

A line in the chat widget header that is visible before or at the first message.
A first bot message that says it is an AI.
Both of the above. This is the safest default, and it costs one setting and one sentence.

Placements that fail

Disclosure only in the site footer.
Disclosure only in the terms of service or privacy policy.
Text that is tiny, faint, or gone before anyone can read it.
A robot avatar or a bot-style name in place of words. The "obvious" exception is judged from a reasonable user's point of view, and a cute avatar is a weak thing to bet on.

Wording you can adapt

Pick the tone that fits the brand. Replace the placeholders in square brackets.

Neutral and professional

Widget header: AI assistant
First message: "Hi, I'm an AI assistant and I can answer most questions straight away. Ask me anything, or type 'human' to reach the team."

Warm, for consumer brands

Widget header: Chat, powered by AI
First message: "Hey, I'm [BotName], [Brand]'s AI helper. I'm not a person, but I'm quick. What can I find for you today?"

Minimal, for B2B and developer tools

Widget header: AI chat (automated)
First message: "This is an automated AI assistant. Responses are machine-generated. Type 'agent' to reach a person."

The neutral version in German

Widget header: KI-Assistent
First message: "Hallo, ich bin ein KI-Assistent und beantworte die meisten Fragen sofort. Fragen Sie mich, oder tippen Sie 'Mitarbeiter', um das Team zu erreichen."

Phone and voice agents need the disclosure spoken before the conversation starts: "You're speaking with an AI assistant for [Brand]. Say 'representative' at any time to reach a person."

Match the language of the bot

If the bot answers in German, the disclosure should be in German. A single English header over a German conversation is a weak position if someone challenges it. Wherever a client's bot serves several languages, translate the header and the first message together. Have a native speaker read translations that will face customers, especially for French and Spanish markets, since a stiff or wrong translation reads as careless.

When a human takes over

If the bot can hand a conversation to a person, say so at the point of transfer: "You're now chatting with [name], a member of our team." It ends the AI context cleanly and avoids a user wondering who they are talking to.

Where to change it

Most chat platforms have two settings that do the whole job: a widget title or subtitle, and a welcome or greeting message. Custom builds need the header line in the widget frame, so it stays visible, and a first message sent from the bot at the start of the session.

Keep a record

For each bot, write down the platform, where each line was placed, the date, and take a screenshot of the live widget. If a client or a regulator asks later, that is the evidence.

Sources

Article 50 text at artificialintelligenceact.eu, in particular 50(1) and 50(5)
Regulation (EU) 2024/1689 on EUR-Lex
European Commission page on the Code of Practice on transparency of AI-generated content, which states that the Article 50 requirements are legal obligations applicable from 2 August 2026

Does the EU AI Act apply to chatbots you build for clients?

mdodeja — Tue, 13 Oct 2026 23:29:57 +0000

Short answer. Probably yes, if people in the EU can use the chatbot, wherever you and the client are based. Which duties are yours depends on whether you count as the provider (you built or supplied the system) or the deployer (you run it).

Where the rule reaches

The AI Act is not limited to companies based in the EU. Article 2 extends it to providers and deployers outside the EU when a system is placed on the EU market or when its output is used in the EU. A chatbot on a US client's website that visitors from Germany or France can open and use is a realistic example.

Whether a given site counts as offered in the EU can be a judgment call. If you don't geo-block and the site accepts EU visitors, the safe assumption is that it is in scope.

Provider or deployer?

Article 50 splits its duties between two roles, so the first job is working out which one you are. Article 3 defines them:

The provider develops an AI system, or has one developed, and puts it on the market or into service under its own name or trademark.
The deployer uses an AI system under its own authority in a professional setting.

Here is how that tends to play out for agencies.

What you did	Likely role	Worth knowing
Embedded and configured a vendor's chatbot on a client site, under the vendor's name	Deployer. The vendor is the provider.	You still control the widget text, so check what the vendor shows by default.
Built a bot with your own code on top of a model API and handed it to the client	Possibly the provider of that chatbot	The model company is the provider of the underlying model. The chatbot you built on top can be a separate system with you as its provider.
Took a vendor's white-label bot and sold it under your agency's brand	Possibly a provider	Putting your own name or trademark on a system is part of the definition.
Run a client's existing bot day to day	Deployer, or an operator acting for one	Your contract should say who is responsible for what.

These are common readings, not settled answers. Article 3 gives the definitions, but how they fit agency arrangements is still being worked out. When your role is unclear, treat yourself as responsible for the disclosure.

What Article 50 asks of a chatbot

Article 50(1) says AI systems meant to interact directly with people must be designed so that those people are told they are dealing with an AI system. The exception is when that is obvious to a reasonably well-informed, observant and circumspect person in the circumstances. Article 50(5) adds that the information has to be clear, distinguishable, and given at the latest at the first interaction.

The 50(1) duty sits with the provider. Deployers have their own duties in 50(3) (emotion recognition and biometric categorisation) and 50(4) (deepfakes and certain AI-written text).

So if you only deploy someone else's bot, the legal duty to build disclosure in belongs mainly to the vendor. In practice the widget header and the welcome message are settings you control, and a client's lawyer will ask you why they are missing, not the vendor. That is why most agencies are better off adding the disclosure themselves than waiting on a vendor.

What to do this week

List every chatbot and voice agent you run for a client, and note whether people in the EU can reach it.
For each one, write down your role: provider, deployer, or unclear.
Put the disclosure in the widget header and in the bot's first message. The wording guide has examples.
If a vendor supplies the bot, ask them in writing what they do to meet Article 50(1). Keep the answer.
Take a dated screenshot of the live widget.

Penalties

Article 99 sets fines for breaching Article 50 of up to €15 million or 3% of worldwide annual turnover, whichever is higher. For small and medium-sized businesses the cap is whichever of the two is lower. National authorities enforce the Act and the practice is new, so nobody can say yet how strictly Article 50 will be policed.

Sources

Article 50 text at artificialintelligenceact.eu
Regulation (EU) 2024/1689 on EUR-Lex, the official text. See Article 2 (scope), Article 3 (definitions) and Article 99 (penalties).
European Commission page on the Code of Practice on transparency of AI-generated content, which confirms that the Article 50 transparency requirements apply from 2 August 2026.