Air Canada Chatbot Ruling: Who Is Liable When Your AI Assistant Gets It Wrong

The short answer

You are. In the Air Canada chatbot ruling, Moffatt v. Air Canada, 2024 BCCRT 149, British Columbia's Civil Resolution Tribunal held the airline responsible for wrong refund advice its website chatbot gave a grieving customer, rejecting the argument that the bot was somehow a separate entity. If your AI assistant tells a customer something false and they reasonably rely on it, expect the bot's words to be treated as your words. This is general information, not legal advice.

What happened in the Air Canada chatbot ruling

After his grandmother died, Jake Moffatt used the chatbot on Air Canada's website while booking a flight to Ontario. According to a summary by Torkin Manes, the bot told him bereavement fares were available and that he could apply for the reduced rate within 90 days of the ticket being issued. That was wrong. Air Canada's actual bereavement policy page, which the chatbot linked to, said the policy did not apply to requests made after travel.

Moffatt booked at full fare, then applied for the partial refund. Air Canada refused. He took it to the tribunal, which issued its decision on 14 February 2024.

The argument that did not work

Air Canada's defence is the part everyone remembers. As Barry Sookman of McCarthy Tétrault summarises it, the airline argued it could not be held liable for information provided by its chatbot, which it described as a separate entity responsible for its own actions. It also argued that the correct information was available elsewhere on its website.

The tribunal disagreed on both counts. Quoting the decision, Sookman's blog records the key lines:

• "While a chatbot has an interactive component, it is still just a part of Air Canada's website."
• "It should be obvious to Air Canada that it is responsible for all the information on its website."
• The tribunal saw no reason why customers should have to double-check information found in one part of a website against another part.

The tribunal found negligent misrepresentation. Air Canada owed Moffatt a duty of care as a service provider dealing with a consumer, the chatbot's statement was inaccurate, the airline "did not take reasonable care to ensure its chatbot was accurate", and Moffatt reasonably relied on it. He was awarded roughly $650 in damages, the fare difference, plus pre-judgment interest and tribunal fees.

Why a small claim matters to every business with a bot

The money was modest. The principle is not. Tribunal decisions do not bind other courts, and Sookman calls the ruling an incremental application of existing law rather than a new doctrine. But that is exactly why it matters: ordinary rules about misrepresentation already cover your AI. There is no "the AI said it" exception.

Lawmakers are writing that principle down. Utah's SB 226 (2025) states that it is not a defence to a consumer protection violation that generative AI made the statement or took the action. Canada's privacy commissioner, in its generative AI principles, puts it plainly: accountability for decisions rests with the organisation, not the tool.

Seven lessons for any business with an AI assistant

1. Your bot speaks for you. Whether it is a website AI chatbot or a phone agent, write your internal rules as if every bot reply were sent by a staff member, because legally it may as well be.
2. Ground answers in your real policies. A bot that answers from your approved documents, using retrieval rather than its general training, is far less likely to invent a refund window. Our explainer on RAG for business covers how this works.
3. Keep policy text in one place. Moffatt's bot contradicted a page on the same site. If the bot and the website draw from the same source, they cannot disagree.
4. Fence off high-stakes topics. Refunds, pricing exceptions, eligibility, legal or medical questions: have the bot quote the policy and route to a human rather than improvise.
5. Test before launch and after every change. Build a set of tricky questions, including ones where the honest answer is "no", and rerun it whenever the model or your policies change.
6. Keep transcripts. If a customer disputes what the bot said, you need the record. So does your quality team.
7. Do not count on disclaimers alone. Sookman notes Air Canada did not appear to rely on website terms disclaiming liability. A small-print disclaimer that contradicts what the bot confidently says is unlikely to be a reliable shield, and fixing the bot is cheaper than litigating its wording.

Many of these are design choices, not legal ones. We go further into them in AI customer service that works.

The commercial maths

Air Canada spent time and reputation fighting a claim worth a few hundred dollars, and the case became a widely discussed example of a chatbot going wrong. For a small business the calculation is even clearer: when your bot gets something wrong in the customer's favour, honouring it quickly is usually cheaper than arguing, and then you fix the bot.

Key takeaways

• In the Air Canada chatbot ruling, the tribunal treated the chatbot as part of the company's website, so the company owned its answers.
• The legal basis was ordinary negligent misrepresentation; no special AI law was needed.
• Utah has since written "the AI did it is no defence" into statute, and Canada's privacy regulator says accountability stays with the organisation.
• Grounded answers, human handoff for money questions, testing and logs are the practical defence.

Frequently asked questions

Is a company legally responsible for what its chatbot says?

Generally, yes. In Moffatt v. Air Canada the tribunal held the airline responsible for all information on its website, including chatbot answers, and found it negligent for not taking reasonable care that the bot was accurate.

Was Moffatt v. Air Canada decided by a court?

It was decided by British Columbia's Civil Resolution Tribunal, an online tribunal for small claims. Its decisions do not bind courts, but the reasoning applies well-established negligence principles that courts also use.

Can a disclaimer protect me from chatbot mistakes?

A disclaimer may help at the margins, but Air Canada did not rely on one and the tribunal said customers should not have to cross-check the bot against other pages. Accurate answers and a clear route to a human are a much better defence.

Sources

1. Moffatt v. Air Canada: A Misrepresentation by an AI Chatbot — Barry B. Sookman, McCarthy Tétrault TechLex, 19 February 2024
2. Moffatt v. Air Canada: A Misrepresentation by an AI Chatbot — barrysookman.com, 16 February 2024
3. BC Tribunal Confirms Companies Remain Liable For AI Chatbot-Created Information — Torkin Manes LLP via Mondaq, 7 March 2024
4. S.B. 226 Artificial Intelligence Consumer Protection Amendments (Enrolled Copy) — Utah State Legislature, 2025
5. Principles for responsible, trustworthy and privacy-protective generative AI technologies — Office of the Privacy Commissioner of Canada, 7 December 2023

How Flowmatix can help

We put this into practice for businesses in Toronto and around the world through our AI Agents & AI Solutions, ai receptionist: chat & phone, software testing & qa services. Want to know what it would take for your business? Get a personalised quote in about 3 minutes.