Skip to content
Technology, Privacy & AI

AI Chatbots and Liability After Moffatt v. Air Canada

Businesses are responsible for what their AI chatbots tell customers. In Moffatt v. Air Canada, 2024 BCCRT 149, the tribunal rejected the idea that a chatbot is a separate entity and found negligent misrepresentation. Consumer protection and competition law also apply, disclaimers have limits, and Bill C-34 would regulate certain companion-style chatbot services.

Customer chatting with an airline website assistant, illustrating liability for AI chatbots

Key takeaways

  • Moffatt found an airline liable for negligent misrepresentation by its website chatbot.
  • The tribunal said a business is responsible for all information on its website, static or chatbot.
  • Customers need not cross-check a chatbot against other pages of the same site.
  • Competition Act s. 74.01 applies to misleading representations by any means.
  • Bill C-34 would regulate chatbot services that simulate sustained human-like relationships.
  • Grounding, scope limits, testing and human hand-offs reduce liability risk.

AI chatbots and liability after Moffatt

AI chatbots now answer customer questions about prices, policies, refunds, eligibility and products for businesses of every size. When a chatbot gets it wrong, who is responsible? The most cited Canadian answer is Moffatt v. Air Canada, 2024 BCCRT 149, a small claims decision of the British Columbia Civil Resolution Tribunal. The amount at stake was modest, but the reasoning has been widely discussed because it rejects the idea that a business can distance itself from what its chatbot tells customers.

Customer chatting with an airline website assistant, illustrating liability for AI chatbots
A business is responsible for information its customer-facing chatbot provides, as it is for any other part of its website.

This insight reviews the decision, the legal theories that apply to chatbot statements, the limits of disclaimers, the emerging online safety framework for certain chatbot services, and practical controls for customer-facing AI. It is general information, not legal advice.

The facts in Moffatt v. Air Canada

In November 2022, following the death of a grandparent, a customer used a chatbot on the airline's website while researching flights. The chatbot suggested that the customer could apply for bereavement fares retroactively. The words "bereavement fares" in the chatbot's answer were a hyperlink to a separate webpage on the airline's bereavement policy, which said the policy did not apply to requests made after travel had been completed. The customer booked flights and later applied for a partial refund, which the airline refused because retroactive applications were not permitted. An airline representative later acknowledged that the chatbot had provided "misleading words" and said the issue had been noted so the chatbot could be updated.

The customer claimed the difference between the regular fares paid and the alleged bereavement fares. The airline argued that the customer had not followed the proper procedure, that it could not be held liable for information provided by the chatbot, and that terms of its tariff limited its liability.

The decision

Tribunal Member Christopher C. Rivers, in a decision dated 14 February 2024, mostly allowed the claim on the basis of negligent misrepresentation. The tribunal found that, given the commercial relationship between a service provider and a consumer, the airline owed the customer a duty of care, and that the applicable standard required it to take reasonable care to ensure its representations were accurate and not misleading.

The tribunal described as "remarkable" the airline's suggestion that the chatbot was effectively a separate legal entity responsible for its own actions.

While a chatbot has an interactive component, the tribunal said, it is still just part of the airline's website, and it should be obvious that the company is responsible for all the information on its website, whether it comes from a static page or a chatbot. The tribunal found that the airline did not take reasonable care to ensure its chatbot was accurate, and that the customer's reliance was reasonable: there was no reason a customer should know that one part of the website was accurate and another was not, or should have to double-check information found in one part of the website against another.

The tariff argument failed because the airline did not provide the relevant portions of the tariff. The tribunal ordered payment of $650.88 in damages, $36.14 in pre-judgment interest and $125 in tribunal fees, a total of $812.02.

What the decision does and does not decide

As a small claims decision of an administrative tribunal, Moffatt is not binding on courts, and its facts were simple. It did not consider a carefully drafted, properly evidenced limitation of liability, a business customer, or a chatbot clearly presented as experimental. But its core reasoning reflects ordinary legal principles that courts are likely to apply: a business that deploys AI chatbots to communicate with customers adopts what they say, and the fact that an answer was generated by software rather than typed by an employee does not change the analysis. The decision also signals that tribunals will not readily expect consumers to cross-check a chatbot against other pages of the same website.

TheoryKey elementsChatbot application
Negligent misrepresentationDuty of care, untrue or misleading statement, negligence, reasonable reliance, damagesThe basis of Moffatt; inaccurate answers about policies or prices
ContractStatements may form part of, or vary, the contractChatbot promises about terms, refunds or features
Consumer protection statutesProhibitions on deceptive or unfair practicesMisleading answers to consumers; statutory remedies
Competition ActFalse or misleading representations to the public in a material respectChatbot claims promoting products or business interests
Privacy lawConsent, purpose limitation, safeguardsCollection and use of personal information in conversations

Consumer protection and competition law

Beyond negligence, statements by AI chatbots can engage statutory regimes. Under section 74.01(1)(a) of the Competition Act, a person engages in reviewable conduct if, to promote a product or business interest, it makes a representation to the public that is false or misleading in a material respect, by any means whatever. Nothing in that wording exempts automated communications.

Provincial consumer protection statutes, such as British Columbia's Business Practices and Consumer Protection Act, prohibit deceptive acts or practices in consumer transactions and provide remedies that can include damages and the ability to cancel agreements. For businesses, the practical point is that chatbot outputs should be treated as marketing and customer communications subject to the same legal standards as any other.

The limits of disclaimers

Many businesses add disclaimers stating that chatbot answers may be inaccurate and that customers should check official policies. Disclaimers can help, but their effect is limited. In Moffatt, the tribunal was unpersuaded that customers should have to double-check one part of a website against another. Consumer protection law can override or limit the effect of disclaimers in consumer contracts, and a disclaimer that contradicts the overall impression created by a chatbot presented as a reliable source of help may not prevent a finding that a representation was misleading.

Disclaimers are more likely to be effective when they are prominent, specific, consistent with how the chatbot is presented, and backed by design choices, such as linking to authoritative sources and routing high-stakes questions to people.

Practical point

Do not rely on a disclaimer to fix an inaccurate chatbot. Restrict the chatbot to approved sources for policies, prices and eligibility rules, test it against common questions, and route refund, cancellation and other high-stakes requests to a human or to authoritative pages.

When AI chatbots become regulated services

Most customer-service chatbots are not specifically regulated today. That may change for some services. Bill C-34, a government bill introduced on 10 June 2026 to enact the Digital Safety Act and establish a Digital Safety Commission of Canada, would regulate "chatbot services" among other online services. The bill defines a chatbot service as an AI system publicly accessible in Canada that uses a natural language interface to provide adaptive, human-like conversational responses, is capable of simulating a sustained human-like relationship with a user, such as friendship, an intimate relationship or therapeutic support, and generates responses that are not fully predetermined.

Only chatbot services that meet user thresholds set by regulation, or that are designated because they pose a significant risk of harm, would be regulated.

Operators of regulated chatbot services would have a duty to act responsibly, including measures to mitigate the risk that the chatbot communicates harmful content, measures to ensure intervention in crisis situations and measures to mitigate certain harmful behaviours, along with duties to protect children and to be transparent. As of November 2026, the bill had passed second reading and was at committee in the House of Commons. A typical customer-service bot is unlikely to meet the "sustained human-like relationship" element, but companion and wellness chatbots could. See our insight on online safety legislation.

Privacy obligations in chatbot conversations

Customers routinely share personal information with chatbots, sometimes sensitive information about health, finances or family circumstances, as in Moffatt. Businesses must identify the purposes for collecting conversation data, limit retention, protect it with appropriate safeguards and make sure their AI vendor does not use it for unauthorized purposes. Quebec requires organizations to inform individuals when a decision about them is based exclusively on automated processing. See our insight on AI vendor contracts.

Controls for customer-facing AI chatbots

Businesses deploying AI chatbots can reduce liability risk with a set of practical controls. Ground answers in approved, current content using retrieval from authoritative sources rather than open-ended generation. Define topics the chatbot may not answer, or must hand off to a person, such as refunds, legal rights, medical issues and complaints.

Test AI chatbots before launch and continuously after, with a library of real customer questions and edge cases. Monitor conversations for errors and complaints, and fix sources quickly when policies change. Keep logs so that disputes can be investigated, subject to privacy retention limits. Identify the chatbot clearly as AI. Align the chatbot with terms of service, and make sure frontline staff know how to handle a customer who relied on a chatbot answer.

Handling chatbot errors when they happen

When a chatbot gives a wrong answer that a customer relied on, the cost of honouring the answer is often small compared with the cost of a dispute, a regulatory complaint or negative publicity. Businesses should have a policy on when to honour chatbot statements, document the incident, correct the underlying content and consider whether other customers received the same answer. Systematic errors may need broader remediation, including proactive outreach to affected customers and a review of how the faulty content entered the chatbot's sources.

Sector-specific considerations

Regulated industries face additional constraints on AI chatbots. Financial institutions must meet consumer protection and complaint-handling obligations under federal financial consumer protection rules and their regulators' expectations for third-party and model risk. Insurers and investment firms must avoid giving advice that requires registration or a suitability assessment. Health care providers must avoid chatbot outputs that amount to clinical advice without appropriate oversight and must protect health information under health privacy statutes. Telecommunications and utilities face sector codes on billing and disclosure. In each case, the chatbot's scope should be mapped against what the business is permitted, and required, to say.

Governance and accountability

Customer-facing chatbots should sit within the organization's AI governance program, with a named business owner, a risk assessment, approval before launch and periodic review. Legal, compliance and customer service teams should agree on the scope of what the bot may say and on escalation paths. Vendor contracts should address accuracy, change management, logging and incident cooperation.

How TRW & Co helps

We advise businesses deploying AI chatbots on liability risk, consumer protection and competition law compliance, disclaimers and terms of service, privacy, vendor contracts and the emerging online safety framework, and we help respond to customer disputes and regulatory inquiries. See our consumer protection and artificial intelligence pages or contact us.

Frequently asked questions

What did Moffatt v. Air Canada decide?

The BC Civil Resolution Tribunal held the airline liable for negligent misrepresentation after its website chatbot wrongly told a customer they could apply for a bereavement fare retroactively. It awarded $650.88 in damages plus interest and fees, totalling $812.02.

Is Moffatt binding on courts?

No. It is a small claims tribunal decision. But its reasoning reflects general principles of negligent misrepresentation that courts are likely to apply to chatbot statements.

Will a disclaimer protect us?

Disclaimers help only to a limited extent, especially with consumers. Prominent, specific disclaimers combined with design controls and human hand-offs for high-stakes questions are more effective.

Are chatbots regulated by online safety law?

Bill C-34, introduced in June 2026, would regulate chatbot services that simulate sustained human-like relationships and meet user thresholds or are designated. It was at committee in the House of Commons as of November 2026.

What controls should we put in place?

Ground answers in approved sources, limit the topics the chatbot handles, test and monitor continuously, log conversations within privacy limits, identify the chatbot as AI and route high-stakes requests to people.

Sources

  1. Moffatt v. Air Canada, 2024 BCCRT 149 (Civil Resolution Tribunal)
  2. Competition Act, s. 74.01 (Justice Laws)
  3. Bill C-34 (45-1), first reading (Parliament of Canada)

General information only, not legal advice. References to statutes, courts and authorities describe published law and practice at the date shown.