Artificial Intelligence: Consequences and a Simple Solution
As one cautiously wades into the ocean of the potential of artificial intelligence, the prudent practitioner must be mindful not to find themselves in hot water.
It’s been a over a year since I wrote about Justice Myers’ cutting commentary in Ko v. Li, 2025 ONSC 2766, where he raised concerns about “artificial intelligence hallucinations” and took the time to list some of a lawyer’s embarrassingly obvious duties, such as the duty not to fabricate case precedents and not to miscite cases.
We continue to learn from the wake of Ko v. Li. Earlier this year, the recipient of Justice Myers’ much-needed commentary found herself in front of the Law Society Tribunal. In Law Society of Ontario v. Lee, 2026 ONLSTH 136, the three-person panel issued a six-month suspension and ordered $10,000 in costs. The misconduct of submitting a factum that relied on AI hallucination of non-existent case law was aggravated by the lawyer’s subsequent dishonesty to the court.
We may find ourselves thinking that hallucinating a case is so blatantly negligent. However, some missteps are more inconspicuous. In Kapahi Real Estate Inc. v. Elite Real Estate Club of Toronto Inc., 2026 ONSC 1438, Justice Myers, again taking up the mantle of a guardian of integrity for civil procedure, deals with what he referred to as the next generation of AI hallucinations.
In Kapahi, counsel delivered a factum citing real cases – so far, so good – but then added fake quotations. Counsel’s position was that these were human errors. He denied that he used AI. Justice Myers noted that the most obvious explanation for the fake quotations is that counsel used AI to draft the factum. However, he did not make that finding, as he did not have the benefit of full submissions on that issue.
Counsel accepted responsibility for his “misreading the cases cited”, “carelessness”, and “inadvertence”. However, Justice Myers noted that, if AI was not used, then he was at a loss for how seven paragraphs of quotations could be a result of human error, a lack of due care, misreading the cases cited, carelessness, or inadvertence.
Justice Myers identified bodies, like the Toronto Police Service and the Law Society of Ontario, as equipped to investigate wrongdoing if they exercise the discretion to do so. He noted that the best outcome to determine if there has been either use of AI hallucinations or deliberate falsification of law is for those with the authority to investigate to be left to do so. He also noted the limitation in his role as judge to investigate, cross-examine or challenge credibility.
In Ko v. Li, Justice Myers was able to refer the contempt of court proceeding against that lawyer to the Attorney General, because counsel admitted in writing that she had been untruthful in a prior hearing. In Kapahi, the facts surrounding the drafting of the impugned factum were unclear, so Justice Myers could not address his concerns in the same way, though he did refer the decision to the Law Society of Ontario for its consideration.
Again, we see the consequences of AI misuse in Law Society of Ontario v. Mazaheri, 2026 ONLSTH 194. There have been several decisions related to this matter starting in 2024, when Mr. Mazaheri had his legal licence suspended on an interlocutory basis. He brought a motion to vary or remove the suspension. As part of that motion, he filed AI-generated materials with the Law Society Tribunal that contained AI hallucinations or fabrications. The Tribunal ordered the lawyer to pay $31,150 in full costs. In the October 20, 2026, decision, the Tribunal held that,
The Lawyer recklessly used a technology tool he knew nothing about, failed to review that technology’s output as he should have, and then filed the resulting error-filled materials with the Tribunal. In doing so he breached his professional obligations of honour and integrity.
The complexity of the underlying technology is mindboggling. The challenges of regulating AI use and integration is potentially just as complex. We see the regulatory struggles facing many fields across society, including academic institutions, artists, and just about anyone who makes a living typing on a keyboard. In law, in Ontario, we are learning from cases like the ones above and starting to see guidance with practice directives and even changes to the Rules of Civil Procedure.
In 2026, Ontario civil litigation has been shaped by themes that included tighter accountability for AI use and greater scrutiny of digital evidence. The caselaw suggests that Ontario courts are treating AI as a tool subject to human certification and judicial supervision, rather than as a basis for an outright prohibition. Artificial intelligence can help with legal research and drafting, but it does not relieve lawyers or litigants of responsibility for what they put before the court. Though the regulatory solutions may be complex, the guiding principle for AI use appears to be straight-forward: check you work.