AI

BMW Dealership Revokes AI Chatbot's $27K Car Buyback Offer

A Toronto customer's experience highlights the legal and operational risks companies face when deploying AI without clear guardrails.

Omega Editorial· June 11, 2026· 3 min read

When the deal-maker isn't human

A BMW dealership in Toronto initially refused to honor a $27,162.79 vehicle buyback offer made by its AI chatbot, telling the customer the bot had made a mistake. The dealership reversed course only after CBC News contacted them about the dispute.

Zack Giacomelli, 31, submitted an online inquiry last month to sell his 2021 BMW back to the dealership where he'd purchased it. He exchanged texts with "Quinn," who asked questions about the vehicle and eventually made a firm buyback offer that would cover what Giacomelli still owed on the car.

Minutes after receiving the offer, a human sales consultant called to revoke it, explaining Quinn was an AI chatbot that had generated the figure in error. The dealership said the actual offer would be around $20,000—more than $7,000 less than Quinn's promise.

According to BMW Toronto sales manager Scott Shadbolt, the chatbot misinterpreted information about how much Giacomelli owed on the vehicle, treating that amount as the buyback price. "The AI bot ran with that, misunderstood the message," Shadbolt told CBC News.

Why it matters

This incident arrives as Canadian businesses rapidly scale AI adoption—12 percent now use AI to produce goods or deliver services, double the rate from a year earlier, according to Statistics Canada. Among those companies, a quarter deploy chatbots or virtual agents. Each implementation carries contractual risk: Canadian courts have already established that companies remain liable for their AI tools' mistakes, treating chatbots like employees for legal purposes.

Legal precedent already exists

Canadian law provides clear guidance on AI liability. In 2024, the British Columbia Civil Resolution Tribunal ruled that Air Canada must honor a fare rebate its chatbot incorrectly promised, rejecting the airline's argument that the bot was "a separate legal entity."

"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. "It can enter into a contract on your behalf."

Walker noted that Giacomelli could have argued for a binding agreement even though he hadn't formally accepted Quinn's offer—the chatbot had scheduled an in-person meeting "to lock in" the deal, which could constitute acceptance.

Transparency failures compound the problem

Giacomelli told CBC News he had no idea he was communicating with AI. Quinn never disclosed its non-human nature during their text exchange. "I feel embarrassed, and I feel angry that I've been negotiating with this bot," he said.

Shadbolt acknowledged the dealership is now implementing changes to ensure customers know when they're interacting with AI, and that only human employees will present buyback offers going forward. "It's a bit of a new territory for us," he said.

Gus Skorburg, co-director of the Centre for Advancing Responsible and Ethical Artificial Intelligence at the University of Guelph, said companies deploying AI to cut costs must accept responsibility when those tools fail. "When they make mistakes, you have to be willing to sort of own up to that," he told CBC News.

These details were first reported by CBC News.

#ai chatbots#contract law#customer service ai#business liability#automotive retail#ai transparency

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in AI

AI· 3 min read

Anthropic, OpenAI Lead U.S. AI Firms' London Office Expansion

American tech giants are racing to secure space in the U.K. capital, drawn by deep AI talent pools but facing infrastructure constraints.

Via AI Watch · Jun 11, 2026
AI· 2 min read

Deezer Opens AI Music Detection Tool to All Streaming Services

The French platform scans playlists from Spotify, Apple Music, and 18 other services after competitors declined to license its technology.

Via AI Watch · Jun 11, 2026
AI· 4 min read

Amazon Teams Hit 10x Productivity With AI-Native Workflows

Six engineers completed an 18-month project in 76 days by restructuring work around AI agents, not just using them as coding assistants.

Via AI Watch · Jun 11, 2026