Can a company escape chatbot liability by calling it separate?

In Moffatt v. Air Canada, 2024 BCCRT 149 (decided 2024-02-14), the BC Civil Resolution Tribunal held Air Canada liable for negligent misrepresentation after its website chatbot told a customer, Jake Moffatt, that he could apply for a bereavement fare retroactively — advice contradicted by another page on the airline's own site. The tribunal found "Air Canada owed Mr. Moffatt a duty of care" given "their commercial relationship as a service provider and consumer," and that the airline breached it: "the applicable standard of care requires a company to take reasonable care to ensure their representations are accurate and not misleading," and Air Canada "failed to exercise reasonable care to ensure the information's accuracy." It also found Moffatt's reliance reasonable and his resulting loss compensable.

The reasoning turns on rejecting Air Canada's defense that the chatbot was a "separate entity" responsible for its own output. The tribunal held that "it is responsible for all the information on its website... It makes no difference whether the information comes from a static page or a chatbot." The McCarthy Tétrault summary situates this within an older tort doctrine, quoting Sookman: "an electronic agent, such as a computer program or other automated means employed by a person, is a tool of that person... the employer of a tool is responsible for the results obtained by the use of the tool since the tool has not independent volition of its own," and a parallel US holding (State Farm v. Bochorsf) that "a computer operates only in accordance with the information and direction supplied by its human programmers. If the computer does not think like a man, it is man's fault."

This is a sharp contrast to Does the UN panel misframe the OpenAI breach as alignment?: where that UN panel brief recasts a corporate AI failure as a technical alignment problem and sidelines the operator's liability, this tribunal does the opposite — it treats the question of why the chatbot erred as irrelevant to who answers for it, attributing the output straight to the operating company under ordinary tool-liability doctrine. It is also a concrete, already-decided instance of the allocation that Can three-tier AI oversight actually prevent deployed system harms? proposes prospectively: a company bearing responsibility for its deployed AI system's output to the public.

The excerpt itself flags the limits of the precedent: the summary notes "Air Canada did not provide any information about the nature of its chatbot," so the ruling says nothing about whether evidence of careful training and testing would have changed the negligence finding, and the author expects "future cases will focus carefully on this." The case also predates and does not address generative AI chatbots specifically, or liability among the "numerous possible AI actors and users" once a system is not simply the operator's own pre-programmed tool. The implication the evidence supports is narrow: a company cannot disclaim liability for its own customer-facing chatbot by calling it separate from itself, but how liability apportions once third-party models, fine-tuning, and multiple AI actors are involved remains open.