I sell into the EU and run an AI chat on my store. Does the AI Act reach me?
AI Act Article 50 transparency duties apply from August 2, 2026 and reach a provider outside the EU where the output is used there. Read your own chat first.
- When this applies
- Your storefront runs an AI chat or shopping assistant, or publishes AI-generated images or copy, and customers in the EU can reach it. [1]
- When this does not apply
- The chatbot disclosure duty in Article 50(1) is written for the provider of the AI system, meaning whoever develops it or has it developed and places it on the EU market or puts it into service under their own name or trademark. Which side of that line your store sits on is a question for your own legal advice. [1]
- Why this matters
- The obligations applied from August 2, 2026, they reach providers established outside the EU where the output of the system is used in the EU, and enforcement carries fines of up to 15 million euros or 3% of total worldwide turnover. [1]
- What to do, or not do
- Open your storefront signed out, start the chat, and read what its first message tells a customer. Then ask whoever supplies that chat who they consider the provider of the system and what disclosure they rely on. [1]
- What you'll see when it's done
- You can name the supplier of every AI system a customer can interact with on your store, and you have read for yourself what each one says at the start of a conversation. [1]
- What this result does not prove
- Reading your own chat proves what it said on the day you read it. It does not establish who the law treats as the provider in your case, and nothing here is legal advice. [1]
- Sources checked
- September 11, 2026
- Changed
- August 2, 2026
When a third party is worth paying for, and what you do not have to buy
- When paid help is worth it
- Paid help is worth it for the one question a blog cannot answer, which is whether your business is the provider or the deployer of the assistant you run, because that answer decides which duties are yours. [1]
Article 50 of the AI Act applies from August 2, 2026. The European Commission’s FAQ on those transparency obligations says so plainly, and it says that from that date providers and deployers of AI systems must comply with the obligations in that provision. If a shopper in the EU can open a chat window on your store, the first thing to look at is what that chat says in its opening message.
Whose duty it is
The FAQ splits the obligations by role, and the split decides what you are reading for.
A provider is a person or body that develops an AI system, or has one developed, and places it on the EU market or puts it into service under their own name or trademark. Article 50(1) is a provider duty: systems that interact directly with people must be designed so that the people concerned are informed that they are interacting with an AI system, unless that is obvious.
A deployer is a person or body using an AI system under its own authority in a professional capacity. The FAQ lists the deployer duties under Article 50 as informing people exposed to emotion recognition or biometric categorisation systems, disclosing deepfakes, and labelling AI-generated text published to inform the public on matters of public interest.
A store that embeds a vendor’s chat widget, under the vendor’s name, is using an AI system. A store that has an assistant built and puts it into service under its own name or trademark meets the FAQ’s description of a provider. Which of those describes your arrangement is a legal question about your contract, and it belongs to a lawyer rather than to a guidance page.
What the disclosure has to do
The FAQ sets out four cumulative criteria for the interaction duty. The system has to qualify as an AI system, be designed for a genuine two-way exchange with people rather than only collecting data or returning automated responses, communicate directly rather than through a human intermediary, and interact with natural persons. Systems running only in the background, machine to machine, sit outside it.
Where it applies, people must be notified that they are interacting with an AI system from the start of the first interaction, in a clear and distinguishable manner and in accordance with accessibility requirements. The obvious-interaction exception exists, assessed against an average person who is reasonably well informed, circumspect and observant, and the FAQ states that it should be interpreted in a restrictive manner because it deprives people of transparency.
An address outside the EU does not settle it
The FAQ states that providers established or located outside the EU are also subject to the AI Act if the output of their AI system is used in the EU. A store registered elsewhere that sells to EU customers is inside the text on the same terms as one registered in Dublin.
The one date after August 2, 2026
A limited grace period exists, and it is narrower than most summaries of it. It covers only AI systems placed on the market before August 2, 2026, and only the marking and detection obligation for AI-generated content under Article 50(2). Providers of those systems comply from December 2, 2026. Content generated before August 2, 2026 does not have to be labelled retroactively, although the Commission encourages it where possible.
If your product photography or product copy came out of a generative system, that marking duty sits with the provider of the system that generated it. Ask the vendor what it does and get the answer in writing.
What the timetable did and did not move
The Commission’s regulatory framework page records that the AI Omnibus entered into force on July 27, 2026 and reset the timetable for high-risk systems, moving Annex III use cases to December 2, 2027 and systems embedded in regulated products to August 2, 2028. The same page still dates the transparency rules to August 2026. The high-risk clock moved. This one did not.
Enforcement runs mainly through national market surveillance authorities. Fines reach up to 15 million euros or 3% of total worldwide turnover for the preceding financial year, with proportionality available for small and medium-sized enterprises.
What to look at
- Open your storefront in a signed-out window, start the chat, and read the first message a customer gets.
- Ask the company that supplies the chat who they consider the provider of the system, and what their disclosure says.
- If anything published on the store was generated by an AI system, ask the same question about marking.
That gives you a named supplier for each AI system on the store and a first-hand record of what each one says. It proves what your store said on the day you read it. It does not establish who the law treats as the provider in your case, and nothing here is legal advice.