Artificial intelligence has quietly become part of daily life. It drafts our emails, summarizes our documents, answers our questions, and generates images in seconds. These tools are genuinely useful, and I am an enthusiastic believer in their potential. But for everyday users and the businesses they run, convenience can create real legal exposure. A few recent Canadian examples make the point better than any abstract caution.
First, be careful what you put in. Many people paste sensitive material into AI tools without a second thought: a customer’s contract, an employer’s confidential file, or someone else’s personal details. Depending on the platform, that information may be stored, reviewed, or used to train future models. Canada’s federal and provincial privacy commissioners recently found that the way ChatGPT was trained did not comply with our privacy laws, having gathered vast amounts of personal information without proper consent. If the data you enter belongs to someone else, you may be breaching a duty of confidentiality simply by sharing it.
Second, be careful what you take out, because you own the consequences. The clearest Canadian warning came from a tribunal in British Columbia. After Air Canada’s website chatbot gave a grieving customer wrong information about bereavement fares, the airline was ordered to pay damages for negligent misrepresentation. Air Canada argued that the chatbot was a separate entity responsible for its own words. The tribunal disagreed, holding that a company is responsible for everything on its website, chatbot included. For any business owner deploying an AI assistant, the lesson is simple: when your AI makes a promise, you are bound by it.
Third, ownership is murkier than it appears. AI-generated text and images may not enjoy the same copyright protection as human work, and what you publish may unintentionally reproduce material the model learned elsewhere. A coalition of Canadian news organizations, including the CBC, the Toronto Star, and the Globe and Mail, is now suing OpenAI, and a Vancouver author has launched class actions against several AI developers. Assuming that you fully own an AI output, or that it is free of anyone else’s rights, can prove costly.
Fourth, you remain responsible for what you share. If you post AI-generated content that defames a real person or misleads consumers, the liability attaches to you, not to the tool. The same is true of deepfakes and impersonations.
None of this is a reason to avoid AI. It is a reason to use it with judgment. Treat AI output as a first draft rather than gospel. Be deliberate about what you enter. Read the terms of service before you click accept. And when the stakes are genuinely high, such as a contract, a will, or a regulatory filing, seek advice from a qualified human professional. Used thoughtfully, AI is a remarkable assistant. Used carelessly, it can quickly become a legal liability.
Disclaimer: the above information does not constitute legal advice. We strongly recommend obtaining independent legal advice with respect to any legal issues.
