Air Canada chatbot lawsuit: is a company liable for what its chatbot says?

Quick answer In Moffatt v. Air Canada (2024 BCCRT 149), decided February 14, 2024, British Columbia's Civil Resolution Tribunal held Air Canada liable after its website chatbot wrongly said bereavement fares could be claimed after travel. The tribunal rejected the idea that the chatbot was responsible for its own actions and ordered Air Canada to pay C$812.02.

What happened with the Air Canada chatbot?

November 11, 2022. Jake Moffatt's grandmother has just died in Ontario. That same day, Moffatt goes to Air Canada's website to book a flight from Vancouver to Toronto using the airline's bereavement rates.¹

There is a support chatbot on the site. Moffatt asks it about bereavement fares. The answer is clear and confident:¹

If you need to travel immediately or have already travelled and would like to submit your ticket for a reduced bereavement rate, kindly do so within 90 days of the date your ticket was issued by completing our Ticket Refund Application form.

Air Canada's chatbot, as quoted in Moffatt v. Air Canada, 2024 BCCRT 149

So Moffatt books. A one-way flight to Toronto for $794.98. Five days later, a one-way flight back for $845.38.¹ Moffatt also calls Air Canada and is told the bereavement fare would be about $380 per flight.¹

On November 17, well inside the 90 days the chatbot gave, Moffatt applies for the bereavement fare. Weeks of emails follow. In February 2023, Moffatt sends Air Canada a screenshot of the chatbot's answer.¹

Here is the problem. The words "bereavement fares" in the chatbot's reply linked to an Air Canada page that said the policy does not apply to requests made after travel is completed.¹ The bot and the website said opposite things.

On February 8, 2023, an Air Canada representative admitted the chatbot had provided "misleading words," pointed to the linked page, and said the airline had noted the issue so it could update the chatbot.¹ But no refund. Moffatt took Air Canada to the tribunal.

What did the tribunal decide?

Air Canada's defense is what made this case famous. It argued it could not be held liable for information provided by one of its agents, servants or representatives, including a chatbot.¹ Tribunal member Christopher C. Rivers was not persuaded:

In effect, Air Canada suggests the chatbot is a separate legal entity that is responsible for its own actions. This is a remarkable submission. 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.

Tribunal member Christopher C. Rivers, Moffatt v. Air Canada, 2024 BCCRT 149

"It makes no difference whether the information comes from a static page or a chatbot," the decision adds.¹ The tribunal found Air Canada did not take reasonable care to make sure its chatbot was accurate, and that Moffatt's reliance on it was reasonable: "There is no reason why Mr. Moffatt should know that one section of Air Canada's webpage is accurate, and another is not."¹

Air Canada also argued that the right answer was on another page. The tribunal said the airline did not explain why that page was more trustworthy than its chatbot, or why customers should have to double-check one part of its website against another.¹ The claim succeeded as negligent misrepresentation.¹

The order, issued February 14, 2024, required Air Canada to pay within 14 days: C$650.88 in damages, C$36.14 in pre-judgment interest and C$125 in tribunal fees.¹ The tribunal also refused to subtract a $200 coupon Air Canada had offered, because Moffatt said they never accepted it and Air Canada gave no evidence otherwise.¹ The Guardian reported that the airline came under further criticism for trying to distance itself from the error.²

Why it happened

The decision itself notes that Air Canada did not provide any information about the nature of its chatbot.¹ So we do not know whether it was a scripted bot or a generative AI model, or why it gave the wrong answer. What we know is simpler and more useful: the bot's answer contradicted the airline's own policy page, nobody caught it, and a customer spent real money on the strength of it.

Keep the scale in proportion, too. This was a small claims decision about a few hundred dollars, from a tribunal in one Canadian province. It is not a ruling from a high court. But the reasoning is plain, and it is easy to see why it travelled: a company is responsible for what its website says, whether a person wrote it or a bot did.

What this means for your business

If you run a store with a chatbot or an AI agent answering customers, this case is about you. A few practical steps:

• Treat the bot's words as your words. Refund rules, delivery promises and discounts given by a bot are promises made by your brand.
• Keep one source of truth. The Air Canada bot and the Air Canada policy page disagreed. Make sure your bot answers from the same policy your team uses, and update both together.
• Fence off money topics. Refunds, fees and exceptions are where a wrong answer costs you. Route those to a person, or limit the bot to quoting the exact policy text.
• Review real conversations. Air Canada acknowledged the error only after the customer sent a screenshot, about three months after the chat.¹ Reading a sample of bot chats every week catches this sooner.
• When the bot is wrong, fix it for the customer too. Promising to update the bot did not end this dispute; the customer still wanted the fare they had been told about.

Chatbots are not the only way AI output can embarrass a business. Lawyers learned the same lesson in court in Mata v. Avianca, where ChatGPT invented case law. And for the human check that would have helped here, see why the confirm button matters.

How ZNOS handles this

The way we run it, anything an AI employee says to customers or publishes under your name is treated as risky. In ZNOS, each AI employee owns a job, such as Jason on sales or Raghu on operations, and works from the company's live data through the Brain, so answers come from how your business actually runs.

Autonomy is earned gradually. Routine work is handled, while decisions involving money, customers or anything published under your name wait for a human approval. Content drafted by an agent waits for that approval before it goes out.

The WhatsApp Business Agent hands off to a human when a conversation needs one, and it follows "never-say" phrases you set, so you can keep it away from promises you do not want made. An Audit Log records who changed what. You can meet the ZNOS team, or read more real AI incidents.