Air Canada's chatbot lied to a customer. Then the airline argued in court that the bot was on its own — it didn't go well.
Jake booked a flight after his grandmother died. Before buying, he asked Air Canada's website chatbot about bereavement fares. The bot told him he could book now and apply for the discount within 90 days.
So he did.
Then Air Canada rejected the refund. Because that policy the chatbot described? It didn't exist. The bot made it up.
Jake took them to a tribunal. Air Canada's defense was that the chatbot is a "separate legal entity" responsible for its own actions. In plain English, the airline argued it shouldn't be on the hook for what its own bot said.
The tribunal was not amused. The ruling basically said a chatbot is still just part of your website. You're responsible for everything on it, whether it comes from a page or a bot. Air Canada paid up.
The lesson isn't "don't use chatbots." It's simpler and scarier than that: whatever your AI says, you said. There is no version where the bot is the one who messed up. The customer doesn't see a bot. They see your company making a promise.
A lot of owners are racing to automate customer-facing stuff right now to save a few bucks. Fine. But every automated word that goes out under your name is a word you're legally and morally on the hook for.
If a human employee promised something wrong, you'd own it. The bot is no different. It's just faster at being wrong.
https://www.facebook.com/share/19BJsMwjiz/