In November 2022, a passenger named Moffatt asked Air Canada’s website chatbot about bereavement fares. His grandmother had just died, and he needed to fly from Vancouver to Toronto for the funeral. The chatbot told him he could buy a regular ticket and apply for the discount within 90 days. However, that was wrong. The airline’s actual policy did not allow refunds. In fact, the chatbot had even attached a link to the page that said so.
Air Canada refused the refund. When Moffatt took the case to a tribunal, the airline argued that it was not responsible for what its chatbot said. In February 2024, the tribunal rejected that argument and ordered the airline to pay him the difference.
I think this case highlights one of the core problems with language-model agents. The chatbot gave a confident answer that was completely wrong. There are multiple methods being implemented to mitigate this, and neurosymbolic agents are one response to this problem.…