BMW Dealership Tried to Revoke an AI Chatbot's Car Buyback Offer - Then Had to Honor It Anyway
A Toronto BMW dealership deployed an AI chatbot to handle customer inquiries. The chatbot made a buyback offer on a customer's 2021 BMW that the dealership said it never authorized. The dealership tried to revoke the offer. The customer pushed back, lawyers got involved, and the dealership ultimately reinstated the deal. It is the latest in a pattern of AI chatbot errors that are creating real legal and financial exposure for Canadian businesses - and a legal framework is already in place that puts liability squarely on the companies deploying the tools.
A Toronto man, Zack Giacomelli, was shocked when a BMW dealership revoked an offer to buy back his car, telling him an AI chatbot generated the offer by mistake. The dealership has since reinstated the deal, but as Canadian businesses rush to adopt AI tools, they face a growing risk of customer backlash if those tools make mistakes. Developers Digest
Giacomelli said: "If they're going to be replacing their employees' jobs with AI, then they need to be honouring what that AI says." Developers Digest
The Legal Framework Is Already Set
This case does not happen in a legal vacuum. Canadian courts have already established the principle that companies are liable for what their AI chatbots say to customers.
In a 2024 case, Air Canada was forced to honour a fare rebate after its chatbot provided a passenger with incorrect advice about bereavement fares. The airline argued before the British Columbia Civil Resolution Tribunal that the chatbot was "a separate legal entity that is responsible for its own actions." The tribunal disagreed, stating that Air Canada was ultimately responsible. "Just like an employee may do something wrong and the company's held responsible, a bot is just like an employee," said Tanya Walker, a litigation lawyer with Walker Law in Toronto. "I don't think companies really realize the magnitude and the legal exposure this creates." Developers Digest
The Air Canada precedent is critical context for any business deploying customer-facing AI in Canada. The bot is legally the company. What it says creates obligations.
The Pattern Is Getting Worse
The BMW case follows a string of similar incidents. A Chevrolet dealership in the US became a viral cautionary tale in 2023 after its ChatGPT-powered chatbot was manipulated into "agreeing" to sell a $60,000 SUV for $1. The dealership did not honor the sale, but the reputational damage was severe.
What makes the BMW case different - and more instructive - is that it was not a manipulation attack. The chatbot made a straightforward commercial offer that the dealership apparently had not intended. No external bad actor required. Just an AI system operating without sufficient business logic constraints, making commitments it had no authorization to make.
For business leaders using AI for customer service or AI agents, the practical lessons are immediate. Every customer-facing AI deployment needs clearly defined authority limits - what the system can and cannot offer, commit to, or agree to. Those limits need to be technically enforced, not just written into a terms of service that the AI will never read. And the legal department needs to be involved in the deployment process, not consulted after the chatbot has already created liability.
The Business Cost of Getting It Wrong
Beyond the direct financial cost of honoring an unintended offer, the reputational math is often worse. Giacomelli's frustration - "if they're replacing employees with AI, they need to honor what it says" - captures a sentiment that spreads fast. Customers who watch a company deploy AI to cut costs and then use "the bot made a mistake" as a defense will not stay customers for long.
Canada's legal trajectory is clear. The Air Canada ruling established the principle. The BMW case applies it to a different industry. Further cases will establish it across retail, financial services, insurance, and healthcare. Businesses that build AI deployments with proper authority controls, audit trails, and legal review now will not be the companies reversing customer commitments in public.
Cut Through the Noise
What happened with the BMW dealership AI chatbot in Canada? A Toronto BMW dealership's AI chatbot made a buyback offer on a customer's 2021 BMW that the dealership said was unauthorized. When the dealership tried to revoke the offer, customer Zack Giacomelli pushed back. The dealership ultimately reinstated the deal after legal pressure. The case illustrates the liability risk Canadian businesses face when AI chatbots make commitments without proper authority constraints.
Are companies legally responsible for what their AI chatbots say to customers in Canada? Yes, under Canadian legal precedent. In a 2024 ruling, the British Columbia Civil Resolution Tribunal rejected Air Canada's argument that its AI chatbot was "a separate legal entity responsible for its own actions," ruling that Air Canada was liable for incorrect bereavement fare information its chatbot provided. Canadian courts treat AI chatbots as equivalent to employees - companies are responsible for what they say.
What should businesses do to protect themselves from AI chatbot liability? Businesses should define clear authority limits for every customer-facing AI system - specifying exactly what the AI can and cannot offer, agree to, or commit to. Those limits must be technically enforced in the system's design, not just documented in policy. Legal review should be part of the deployment process. Audit trails of all customer interactions should be maintained to document what was said and under what authority.
How common are AI chatbot mistakes in business deployments? Cases range from unauthorized offers and incorrect information to manipulated agreements and privacy violations. The most publicized include a Chevrolet dealership whose chatbot was tricked into agreeing to sell a $60,000 vehicle for $1, and Air Canada's bereavement fare case that set Canadian legal precedent. As AI chatbot deployments expand across industries, incidents are increasing in frequency and legal significance.
