The Chatbot Made a Promise the Airline Had to Keep
A grieving customer relied on Air Canada's website chatbot for bereavement fare information. When the airline refused to honour what the bot had promised, a Canadian tribunal ruled that the company, not the AI, was responsible.
In November 2022, Jake Moffatt visited Air Canada's website to book a last minute flight to Toronto following the death of his grandmother. He asked the airline's customer service chatbot about bereavement fares, and the bot told him he could book a full price ticket and apply for a partial refund within 90 days. That advice was wrong: Air Canada's actual policy required bereavement discounts to be approved before travel, not claimed afterwards.
Moffatt booked the flight, paid full fare, and later submitted his refund request exactly as the chatbot had instructed. Air Canada refused, telling him the chatbot's answer was inaccurate and non binding. The airline's own staff conceded the bot was at fault and offered a goodwill voucher, but Moffatt chose to take the case to the British Columbia Civil Resolution Tribunal instead.
Air Canada's defence became almost as newsworthy as the original error: the airline argued the chatbot was a separate legal entity, responsible for its own words. The tribunal rejected that argument outright, ruling that a company is liable for everything on its website, whether the information comes from a static page or an interactive chatbot. Air Canada was ordered to pay the promised refund, plus damages and fees.
What went wrong
The case has since become a reference point across industries for a simple principle. Deploying an AI agent to represent a brand does not create a legal buffer between the company and what that agent says. If a chatbot makes a commitment, the business is on the hook for it, whether or not a human ever reviewed the exchange.
Learning outcomes
After discussing this case, participants should be able to:
- Explain why deploying a customer facing AI agent does not shift legal liability away from the company that deployed it.
- Identify the categories of claim, including negligent misrepresentation and consumer protection, that AI generated statements can expose a business to.
- Describe what reasonable care looks like when validating a chatbot's outputs against actual policy documents.
- Assess how quickly your organisation could correct or overrule a chatbot's error before it reaches a tribunal or regulator.
Discussion questions
- If your customer service AI gave a customer incorrect information today, does your organisation have a documented process for honouring or correcting it?
- Who signs off on the source material a customer facing AI draws its answers from, and how often is it audited against current policy?
- Would your legal or risk team be alerted before a chatbot dispute reached a regulator or tribunal, or only after?