Pivotal Decision on Artificial Intelligence from Canadian Tribunal
On February 14, 2024 a Canadian Civil Resolution Tribunal (“Tribunal”) ruled in Moffatt v. Air Canada on artificial intelligence’s personhood and liability in case involving use of artificial intelligence, causing a ripple effect in the international law community.
This all started when, a passenger visited Air Canada’s website to book a flight and was informed by a chat bot that he could receive a retroactive discount on his flight in the event of a bereavement.
The passenger, who purchased his ticket at the price specified on the website based on the chat bot’s instructions that he could receive a retroactive discount, requested a discount within the period specified by the chat bot.
However, the discount conditions stated by the chat bot contradicts the Air Canada policy listed on its website, which explains the conditions for a fare reduction in case of bereavement.
After receiving a response from Air Canada representatives that “the chat bot used misleading words and the problem was noted so that the chat bot could be updated”, the passenger took the matter to the Tribunal on the grounds that Air Canada breached its duty of care.
On the other hand, Air Canada argued that it could not be held liable as the chat bot, which is an artificial intelligence model, has a separate legal personality.
The Tribunal dismissed Air Canada’s claims that the chat bot was a separate legal entity. The Tribunal defined the duty of care as a company taking reasonable care to ensure that its statements are truthful and not misleading, and ruled that as part of Air Canada’s website, any information provided by the chat bot was subject to Air Canada’s duty of care.
It should be underlined that this was a decision by a court of first instance, however it addresses a recurring issue on the legal personality of artificial intelligence and the imputed liability of artificial intelligence users for the actions of artificial intelligence.
We expect that disputes and decisions around this issue will increase, especially with the entry into force of the EU Artificial Intelligence Act.
Successful