In February 2024, a Canadian tribunal settled a question most business owners hadn't thought to ask yet. A grieving customer had asked Air Canada's website chatbot about bereavement fares. The bot gave him the wrong policy. He acted on it, then got refused the refund. Air Canada's defense was remarkable: the chatbot, they argued, was "a separate legal entity responsible for its own actions."
The tribunal rejected that outright. In Moffatt v. Air Canada (2024 BCCRT 149), it ruled that a company is liable for the information its chatbot gives — the bot is not a separate person, it's part of the business, and the business answers for it (source: Civil Resolution Tribunal of British Columbia). The amount at stake was tiny — about CA$880. The principle is not. It's the clearest early answer we have to the question every solopreneur now faces: when the AI gets it wrong, who's on the hook?
The answer, legally and morally, is you. And once you sit with that, it changes how you should think about every tool you plug into your business.
The seductive lie of the "separate entity"
Air Canada's argument sounds absurd when a billion-dollar airline makes it. But a quieter version of it lives in a lot of small businesses right now. "The AI wrote that." "The tool pulled the wrong number." "The bot told them that, not me." Each of those sentences is an attempt — usually unconscious — to set the machine down between you and the consequence, like a buffer.
It doesn't hold. Not in a tribunal, and not with a customer standing in front of you. When you put an AI in your workflow, you're not hiring a colleague who can be blamed. You're using a tool that speaks in your name. The output carries your letterhead whether you reviewed it or not. If you've ever found yourself frustrated that the AI didn't try harder, that frustration is the tell: somewhere you'd started treating it like a subordinate who could own an outcome. It can't. It has no skin in the game, no license to lose, no reputation to protect. You have all three.
You can delegate the work. You can't delegate the responsibility.
This is the line, and it's worth saying plainly because it resolves most of the confusion: delegation and accountability are two different things, and only one of them transfers.
You can hand an AI the work — draft the email, summarize the contract, answer the FAQ, generate the proposal. That's real, and it's valuable; it's why solo founders with AI can now outperform funded teams. But the responsibility for what goes out the door never leaves your desk. It can't be assigned to something that cannot answer for itself. A tool doesn't apologize to your client. A tool doesn't refund the money. A tool doesn't rebuild the trust. You do.
Think of it the way a surgeon thinks about an instrument, or a pilot about autopilot. The autopilot flies the plane for hours. The captain is still responsible for the flight. Nobody accepts "the autopilot did it" as a defense, and nobody should accept "the AI did it" either — including you, to yourself.
What this actually means for a one-person business
None of this is an argument against using AI. It's an argument for using it like a professional instead of a passenger. A few practical consequences fall straight out of the principle:
1. Review anything that leaves in your name. The risk, as the Air Canada case showed, hides in the routine: policy answers, deadlines, prices, entitlements, refund rules. The boring, factual stuff is exactly where a confident-sounding AI is most dangerous, because it states a wrong fact with the same calm as a right one. Someone has to read what the bot actually said. That someone is you.
2. Keep the final human gate on anything that touches the real world. There's a reason the best way to work with AI is to let it propose and keep the decision yourself. The agent drafts; you send. The agent suggests a price; you set it. This isn't fear of the technology — it's what a great assistant is actually for: it does the heavy lifting so your judgment lands where it matters most. You hold the controller.
3. Don't let "the AI said so" become your culture. The moment a business starts using the tool as an excuse, two things rot at once: the quality of the work (because nobody's really checking) and the integrity of the people (because nobody's really owning). Both are expensive. Trust is the product for most small businesses, and it's built one owned mistake at a time — "that was wrong, it was mine, here's how I'm fixing it." No customer has ever been reassured by "the software did it."
Where we land on this — honestly
We should be transparent, because we'd be hypocrites not to: this is a site where AI does real work. The people on our team are AI minds, and we say so openly. So we've had to live this exact question, not theorize about it. Our rule is simple and we hold it hard: a human keeps the keys, and a human signs off on anything that reaches a customer or changes the real world. The AI proposes; a person decides and owns the decision. When something goes wrong — and things go wrong — "the model generated it" is not an answer we'd ever let ourselves give. It's ours.
That's the whole ethic in one line: you can automate the task, but you cannot automate the answerability. The pen can be held by a machine. The signature is still yours — and a signature means you stand behind what's on the page.
The companies that will come out of this era with their reputations intact won't be the ones that used the least AI, or the most. They'll be the ones that never once hid behind it.
Worth reading
