Air Canada's chatbot made a promise. The airline had to keep it.

The case

In November 2022, Jake Moffatt's grandmother died. He needed to fly from Vancouver to Toronto for the funeral. He went to Air Canada's website and asked the chatbot about bereavement fares. ^[1]

The chatbot told him he could buy a ticket at the regular price and apply for the bereavement discount afterwards, within 90 days. That was wrong. Air Canada's bereavement policy, which was on the same website, said the discount had to be requested before travel. ^[2]

Moffatt bought the ticket, flew to the funeral, and then asked for the refund. Air Canada said no. He took the matter to the Civil Resolution Tribunal of British Columbia, a small claims court for disputes up to C$5,000.

In February 2024, the tribunal ordered Air Canada to pay him the difference, plus costs. ^[3]

The airline's argument

Air Canada argued that it could not be held liable for information provided by "agents" such as a chatbot. In effect, the tribunal said, the airline was suggesting the chatbot was "a separate legal entity that is responsible for its own actions".

The tribunal did not accept that. Air Canada is responsible for all the information on its website. It makes no difference whether it comes from a static page or a chatbot. The website as a whole is the airline's.

The airline also pointed out that the correct bereavement policy was published on the website. The tribunal's view was that if there is a contradiction, it is up to Air Canada to make sure there is not one.

The evidence

This is where it gets interesting from a proof point of view.

Moffatt had a screenshot of what the chatbot told him. An Air Canada representative had already admitted, in correspondence before the hearing, that the chatbot's wording was misleading. On the fare itself, the airline offered no evidence to the contrary.

The tribunal said that when one side has evidence and the other does not, it can draw an adverse inference: in other words, assume that if the airline had had helpful evidence, it would have produced it. ^[4]

So the customer's screenshot became the record of what happened. The airline had nothing of its own to show.

What this means for you

If you run a chatbot that talks to customers, the point is fairly simple. What it says on your website is yours. In a dispute, the side without its own records starts on the back foot.

Think of it like a receipt. The shop keeps a copy, the customer keeps a copy. If there is a question later about what was sold, both sides have something to point to. A chatbot that does not keep its own record is a shop that only gives receipts to customers.

Here are some questions worth asking now:

• Could you show what your chatbot said to a particular customer on a particular day?
• Who keeps that record, and can they edit it?
• If a customer screenshots an answer and you disagree, what evidence would you produce?
• Since the EU's AI Act requires you to tell people when they are talking to an AI, are you logging that the notice was given? ^[5]

How a digital witness helps

A digital witness does not stop your chatbot from making a mistake, but it does let you keep your own record of what was said.

Here is how it works in this case. Each time the chatbot sends an answer, your system seals a record: the conversation ID, the customer's question, the chatbot's answer, the timestamp and a note that the AI notice was shown. WitniumChain takes a fingerprint of that record, not the record itself, so the conversation stays in your systems. You get a witness ID back.

Months later, if a customer questions what the chatbot said, you pull up the record and share it with them. They can check that it has not changed by calculating the fingerprint again and comparing it with the sealed one. No trust required: the check is independent.

It does not show that the chatbot's answer was correct, or settle whether you are bound by what it said. What it does is give you evidence of your own, so you are not relying on someone else's screenshot.

How it would work

1. Seal each answer as the chatbot sends it: question, answer, timestamp and AI notice.
2. If it is questioned, share the sealed record with the customer or a tribunal.
3. Anyone can check that the record has not changed since it was sealed.

You can seal through the REST API, the SDK or the MCP server. The content stays with you. The first 2,000 witnesses a month are free. ^[6]

What to watch

Since 2 August 2026, Article 50 of the EU AI Act has required that people be told, in a clear and distinguishable way, when they are interacting with an AI system, unless it is obvious from the context. ^[7] There are narrow exceptions, but the rule applies to most customer-facing chatbots.

That is a transparency rule, not an evidence rule. It says you must inform people. It does not say how you prove you did.

But in a case like Air Canada's, being able to show that the notice was given, and what the chatbot said afterwards, is the kind of detail that matters. Courts care about what actually happened, and records beat assertions.

Sources

1. Moffatt v. Air Canada, 2024 BCCRT 149, Civil Resolution Tribunal of British Columbia, 14 February 2024, https://decisions.civilresolutionbc.ca/crt/crtd/en/525448/1/document.do
2. Air Canada bereavement travel policy, as it stood in November 2022, described in the tribunal's decision.
3. CTV News report: Air Canada's chatbot gave a B.C. man the wrong information. Now the airline has to pay for the mistake, CTV News Vancouver, 16 February 2024, https://bc.ctvnews.ca/air-canada-s-chatbot-gave-a-b-c-man-the-wrong-information-now-the-airline-has-to-pay-for-the-mistake-1.6769454
4. Adverse inference doctrine, applied by the tribunal when one party fails to produce evidence that would reasonably be expected to exist.
5. Article 50 of Regulation (EU) 2024/1689 (AI Act), European Commission AI Act Service Desk, https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50
6. WitniumChain documentation, https://witniumchain.com/docs
7. AI Act implementation timeline, European Commission AI Act Service Desk, https://ai-act-service-desk.ec.europa.eu/en/ai-act/timeline/timeline-implementation-eu-ai-act