Air Canada was held liable for what its chatbot told a customer

What happened

A customer asked Air Canada's chatbot about its bereavement fare policy. The chatbot gave incorrect information, telling the customer they could apply for a refund after the fact. Air Canada refused the refund, citing the airline's actual policy.

The case went to Canada's Civil Resolution Tribunal, which found Air Canada liable for negligent misrepresentation and ordered payment of roughly CAD 812, including CAD 650.88 in damages, interest, and tribunal fees.

The business problem

AI output can be wrong, customers reasonably treat it as official information, and the company remains responsible for the experience either way.

What could have helped

• A company is accountable for what its AI tells customers, full stop
• Set up QA and evaluation for anything customer-facing before launch
• Build a human-escalation path for anything a chatbot can't answer confidently
• Keep chatbot answers in sync with actual current policy, continuously

Where this points