Nothing stops you asking a chatbot a legal question. What it can’t do is take the professional responsibility that turns an answer into advice.
Short answer
No. An AI tool can produce fluent text about a legal question, but “legal advice” in Canada is a service performed by someone licensed to practise law who is personally and professionally accountable for it — a duty an AI system cannot hold. What a chatbot gives you is a starting point for understanding a question, not a substitute for a licensee’s judgment and signature.
As the Law Society of Alberta’s own survey of the field puts it, “since ChatGPT entered the legal lexicon, eleven Canadian courts, nine law societies, two professional liability insurers, one provincial government, the Canadian Judicial Council, Canadian Bar Association and the College of Patent Agents and Trademark Agents have all issued guidance documents regarding the use of Generative AI… They uniformly emphasize that lawyers are responsible for the truth and accuracy of their work and how Gen AI must be used carefully. “None have attempted to ban its use.”” That last sentence is doing the real work: every regulator that has looked at this has left responsibility for accuracy exactly where it already sat — with the licensed person, not the tool.
The Law Society of Ontario’s Futures Committee White Paper, summarised on the same page, sets out a four-factor test for whether a lawyer using Gen AI must tell a client: whether the use will necessarily be disclosed publicly, whether the client reasonably expects a licensee personally prepared the material, whether there is reputational risk to the client, and whether the tool needed the client’s personal or proprietary information. That test was built for lawyers, but the underlying reasoning generalises to any regulated professional deciding whether to say so — see the parallel question for an accountant.
None of that changes the confidentiality exposure that already existed before AI showed up: putting a client’s facts into any third-party tool carelessly is the same kind of problem a confidentiality-clause breach has always been, independent of who or what leaked it.
You can use it to understand the shape of a question. You cannot use it to satisfy a Law Society’s conduct rules, and it carries no professional accountability if it is wrong. Even the Federal Court, describing its own internal use of AI, holds itself to the same discipline it expects of everyone else: “members of the Court and their law clerks are aware of the need to verify the results of any AI-generated outputs that they may be inclined to use in their work” — its own Interim Principles say. If a court will not trust AI output without a human checking it, that caution applies at least as strongly to a legal question about your own affairs. Canada also has no AI-specific statute that changes any of this — see why.
Whenever AI touches a regulated professional’s work, the exposure question is the same one a diligence review would ask: what changed, and who is still accountable for it.
