It’s now beyond all doubt: Artificial Intelligence (AI) is a mind-boggling tool, and staggeringly useful. But in the Canadian legal profession, there’s a serious concern about its misuse – not just by self-represented litigants and lawyers, but also (apparently) by judges too.
Lawyers and Judges and AI – Oh My!
Case in point: Several Ontario lawyers have been caught using AI-driven writing tools like ChatGPT to file inadvertently misleading court materials. Those lawyers have discovered the hard way that their ChatGPT research results produced what are commonly called “hallucinations”, which are fake (but convincing-sounding) legal authorities like cases and legislation.
The Ontario case of Ko v. Li is just one recent example. The lawyer had submitted materials to the court that, unbeknownst to her, referred to hallucinated references to made-up court cases. When called out by the judge, she panicked and blamed her support staff, but later confessed it was a lie. She knew she was the one ultimately responsible for checking the accuracy of any content she filed with the court – including anything derived from ChatGPT. The court reproached the lawyer for the initial misstep, but noted her attempt to deflect responsibility raised new issues about “deception of the court”, and referred the matter to the Attorney General for a contempt of court proceeding.
In another recent case called Kapahi Real Estate Inc. v. Elite Real Estate Club of Toronto Inc., there was a slight twist: The lawyer referenced real cases in his filed court documents, but purported to quote long passages from them that turned out to be wholly made up by AI. To make matters even worse, he then lied to the court about having used AI in his drafting at all. The court referred the matter on to the Law Society of Ontario for its consideration, noting: “As is often the case, if [the accused lawyer] has not been truthful, the cover-up may be worse than the initial error.”
Perhaps even more surprising is the recent scandal involving a Quebec Superior Court judge named Justice Jocelyn Geoffroy. After an initial hearing in a case involving an alleged $120 million Ponzi-scheme and related fraud, one of the litigants accused the judge of using AI in his ruling on the matter. Apparently the judgment referred to non-existent case law, and purported to quote verbatim from evidence and testimony that was not included anywhere in the court records.
The matter is still under investigation, and will likely be referred to the Canadian Judicial Council, which oversees the conduct of judges across the country. Apparently this is the first time a Canadian judge has been accused of incorporating AI-generated content into his or her reasons – and it’s probably not going to be the last.
The Role of Professional Responsibility
Of course the legal professional is not alone in this; the dishonest or unscrupulous use of AI is a problem in many industry sectors. But its misuse by lawyers and judges adds an additional level of concern about ethical and professional standards for individuals, and the integrity of the justice system as a whole.
Fortunately, in Ontario all practicing lawyers are governed by a regulator, the Law Society of Canada, which publishes a detailed code of conduct.
In the next installment in this two-part series, we’ll take a look at how AI use is covered in the those professional standards. We’ll also give you a glimpse at how our firm, Russell Alexander Collaborative Law, addresses our professional duties towards clients .and the court, whenever we do legal research and prepare documents on a file.
For the full text of the decisions, see:
Ko v. Li, 2025 ONSC 6785
Kapahi Real Estate Inc. v. Elite Real Estate Club of Toronto Inc., 2026 ONSC 1438 (CanLII)
Entreprises Bertrand Roberge ltée v. Giroux, 2025 QCCS 4157
