AI chatbot disclosure and human hand-off: what Shopify stores need

Do you have to tell customers they are talking to an AI chatbot?

In several jurisdictions, yes, or close to it. The EU AI Act requires that people be informed when they interact with an AI system, and California prohibits bots that mislead people about being bots to sell goods. Even where no law applies, an AI label is the simplest way to avoid misleading customers.

This article explains the main rules a US-based Shopify merchant is likely to meet: the EU AI Act's transparency article, California's bot disclosure statute and the Federal Trade Commission's position on AI claims. It is a plain-English overview, not legal advice. Your obligations depend on where you sell, who your customers are and how your tools are built, so check with counsel.

Our glossary entry on AI disclosure gives the short definition: telling a person, clearly and up front, that they are dealing with an automated system rather than a human.

What does the EU AI Act Article 50 require for chatbots?

Article 50(1) of the EU AI Act requires providers to design AI systems that interact directly with people so those people are informed they are interacting with an AI, unless that is obvious from the context. Article 50(5) says the information must be clear, distinguishable and given at the latest at the first interaction.

The text is in Regulation (EU) 2024/1689 on EUR-Lex. Article 50(5) also says the information must conform to applicable accessibility requirements, so an AI label that a screen reader cannot announce is a weak label.

Timing matters. The European Commission's AI Act overview states that the transparency rules came into effect in August 2026, and that the AI Act became applicable on 2 August 2026, with some exceptions. The same page notes that the "AI Omnibus" amendments entered into force on 27 July 2026; the changes it lists concern high-risk timelines and other topics, not the chatbot transparency duty.

Who carries the duty? Article 50(1) is addressed to providers, the companies that build the AI system. A merchant who installs a chatbot app is usually using the system rather than building it, but how the Act applies to your setup, and whether you serve EU shoppers in a way that brings you into scope, is a question for counsel. In practice, the safe design choice is the same either way: disclose.

What does California's BOT Act (SB 1001) say?

California's bot disclosure law, enacted by SB 1001 as Business and Professions Code sections 17940 to 17943, makes it unlawful to use a bot to communicate with a person in California online with intent to mislead them about its artificial identity in order to incentivize a sale. Clearly disclosing the bot avoids liability under that section.

Section 17941(b) says the disclosure must be clear, conspicuous and reasonably designed to inform the people the bot communicates with that it is a bot. Section 17940 defines a bot as an automated online account where all or substantially all of its actions or posts are not the result of a person, and also defines a large "online platform" by a monthly US visitor threshold. The law has been operative since July 1, 2019.

Read the statute on the California Legislative Information site, Business and Professions Code section 17941 (an annotated copy is on California Public Law). Note that the prohibition is tied to intent to mislead; a chatbot that announces itself as AI in its first message is designed to avoid that problem. Whether any exception or nuance applies to you is, again, for counsel.

What does the FTC say about AI chatbots and AI claims?

The FTC's position is that existing consumer protection law applies to AI. Businesses must not use AI to deceive people, and claims about what an AI product can do must be truthful and supported. For a store, that covers both how the chatbot presents itself and what you say about it in marketing.

In September 2024 the FTC announced Operation AI Comply, five enforcement actions against companies that overstated AI capabilities or supplied AI tools used to deceive. Then-Chair Lina M. Khan said: "Using AI tools to trick, mislead, or defraud people is illegal." One case involved a service marketed as a "robot lawyer" whose claims, the FTC alleged, were not substantiated.

Two practical lessons follow for a Shopify store. First, the chatbot should not claim abilities it does not have, such as giving professional advice. Second, anything the chatbot says about price, stock, delivery or policy is a claim your store is making, so it has to be accurate. Read more on why chatbots invent facts and how to stop it.

What does a good AI disclosure look like in a store chat?

A good disclosure is visible before the shopper types: the chat window is labelled as an AI assistant, the first message says it is an AI, and the assistant never claims to be human, even when asked directly. It also tells shoppers they can reach a person, and how.

Disclosure fails in quiet ways. A human first name and a stock photo avatar suggest a person, even if a small line of text says otherwise. A label hidden in a tooltip is not conspicuous. A bot that answers "are you real?" with a joke is evasive. Design checks that help:

• Label the window itself, for example "AI assistant", not only the first message, so the label stays visible as the chat scrolls.
• Say it in the first message in plain words: "I'm an AI assistant for this store. I can help with products, sizing and orders, and I can connect you with our team."
• Answer "am I talking to a bot?" honestly every time, in the shopper's language.
• Avoid human names, human photos and typing delays designed to imitate a person.
• Make the label accessible: readable contrast, announced by screen readers, present on mobile.

How should a chatbot hand off to a human?

A good hand-off is easy to ask for, offered without being asked when it is needed, and honest about timing. The shopper should be able to type "talk to a person" at any point, be told who will reply and when, and not have to repeat what they already said.

Our glossary entry on human hand-off defines the term. In practice, design hand-off around the moments where automation should step back:

1. On request. Any phrasing of "human", "agent" or "person" triggers the hand-off, without a retention script.
2. On emotion. An upset or frustrated shopper gets a person, even if the question looks routine.
3. On decisions. Refunds, exceptions, damaged or missing parcels and complaints go to your team.
4. On risk. Medical, legal, financial or safety questions are referred to the label, the merchant or a professional rather than answered.
5. On uncertainty. If the assistant cannot answer from the store's own facts, it says so and offers a person rather than guessing.

What should you avoid saying about your AI chatbot?

Avoid claims you cannot support: that the chatbot is human, that it never makes mistakes, that it is a qualified expert, or that it will produce a specific sales result. The FTC treats AI claims like any other advertising claim. Describe what the tool does, plainly, and keep the human option visible.

This applies to the chatbot's own words as well as your marketing page. An assistant that invents urgency ("only two left!") or social proof that is not real is making a deceptive claim on your behalf. Read how to choose an AI shopping assistant with these checks in your evaluation.

How does Tycoon handle disclosure and hand-off?

Tycoon says plainly that it is an AI, in a window labelled "AI assistant", never pretends to be human, and always offers a person from the merchant's team. Merchants cannot turn this off. Its policies are written with the EU AI Act Article 50 transparency duty in mind, among other frameworks.

It does not give medical, legal, financial or safety advice; it refers to the product label, the merchant or a professional. It does not use false urgency, fake scarcity or fake social proof, and it hands off to the merchant's team when a person is needed. AI disclosure and human hand-off are included on every plan; see security and privacy for how data is handled, or book a demo to see the disclosure and hand-off on your own store.

Frequently asked questions

Sources

1. Regulation (EU) 2024/1689 (AI Act), Article 50, EUR-Lex · checked October 20, 2026
2. European Commission: AI Act, regulatory framework and timeline · checked October 20, 2026
3. California Business and Professions Code §17941 (SB 1001), California Legislative Information · checked October 20, 2026
4. California Public Law: Business and Professions Code §17941 · checked October 20, 2026
5. California Public Law: Business and Professions Code §17940 (definitions) · checked October 20, 2026
6. FTC: FTC Announces Crackdown on Deceptive AI Claims and Schemes (Sept. 25, 2024) · checked October 20, 2026