Is my company liable for what its AI chatbot tells customers?

If your website, app or support desk uses a language model to answer customers, treat its answers as your company’s statements. That’s the clearest lesson from Canada’s most cited chatbot decision.

What happened in Moffatt v. Air Canada?

A customer asked Air Canada’s website chatbot about bereavement fares. The chatbot said he could apply for the discount after travelling, which contradicted the airline’s actual policy on another page. British Columbia’s Civil Resolution Tribunal found Air Canada liable for negligent misrepresentation in Moffatt v. Air Canada, 2024 BCCRT 149. It rejected the argument that the chatbot was a separate entity responsible for its own actions, and said the airline was responsible for all information on its website, whether from a static page or a chatbot.

It’s a small-claims tribunal decision with a modest award and doesn’t bind Ontario courts, but the reasoning follows ordinary negligence and misrepresentation principles that apply across Canada.

Which laws create the risk?

Source of liability	How a chatbot can trigger it
Negligent misrepresentation	A customer reasonably relies on a wrong answer and loses money
Contract	The bot promises a refund, price or feature, and the customer argues it formed part of the deal
Competition Act	False or misleading representations to the public about a product or price
Consumer Protection Act, 2002	Unfair practices, including false or misleading representations, to Ontario consumers
Privacy law	The bot collects or reveals personal information it shouldn’t

Does a disclaimer solve it?

Not on its own. A clear statement that answers are AI-generated and may be wrong helps set expectations, and your terms can limit liability for business users. But a disclaimer buried in terms won’t reliably protect you from a specific wrong answer a consumer relied on, and the Competition Act looks at the general impression of what you said.

How do I reduce the risk?

• Limit scope. Keep the bot to topics it can answer from approved content, and have it decline or hand off on refunds, pricing exceptions, legal, medical or financial questions.
• Ground it. Answer from your current policies and documentation, and link to the source page.
• Escalate. Offer a human for anything involving money, cancellations or complaints.
• Monitor. Log conversations, review samples, and fix recurring errors quickly.
• Disclose. Tell users they’re talking to an AI. See do I have to tell people my product uses AI.
• Contract with your model vendor. Check what its terms say about output, data use and liability. See terms for AI products.

What if I sell a chatbot to other businesses?

Your customers carry this risk with their own users, and they’ll try to pass some of it to you. Your SaaS agreement should say the customer configures and approves the bot’s content, set an acceptable use policy, and keep output-related claims within your limitation of liability.

What to do next

We draft AI product terms and SaaS agreements at fixed fees on our fees page. For the wider picture, see is AI regulated in Canada.

This guide is general information about Ontario and Canadian law, not legal advice for your situation. To talk through an AI feature before launch, book a free 15-minute consult.