Man with suspended licence says he relied on Grok to research, draft documents
The Law Society of Ontario Tribunal’s Hearing Division saw no bias in its response to the issues arising from a lawyer’s use of generative artificial intelligence (AI) in proceedings where he moved to vary or remove the interlocutory suspension of his licence.
A LinkedIn post from Amy Salyzyn, law professor at the University of Ottawa, described this case as “the first reported instance ([at] least in Canada) of a lawyer submitting problematic AI-generated materials in the context of a disciplinary hearing.”
In Mazaheri v Law Society of Ontario, 2025 ONLSTH 186, the applicant was a lawyer whose licence the Law Society Tribunal had suspended on an interlocutory basis.
On Nov. 12, 2024, the applicant moved to vary or remove the interlocutory suspension. In an admissibility and bias motion, he sought to:
- Exclude some evidence filed by the Law Society of Ontario in response to his motion to remove the suspension
- Recuse the tribunal panel members as incapable of unbiased judgment, given that they had reviewed evidence to which he objected
- Recuse the members, as they had criticized him at a case management hearing, interrupted his submissions, and raised professional conduct issues without authority
The tribunal noted that the applicant filed motion materials produced with the assistance of a generative AI tool that was ‘hallucinating.’ The tribunal provided a chart of examples of non-existent and misleading authorities in the materials.
The tribunal directed the parties to appear at a case management conference and address how to proceed in light of the applicant’s materials.
On Nov. 30, 2025, the applicant admitted to the tribunal that he relied on generative AI, specifically Grok, in researching and drafting documents and failed to sufficiently verify his materials, which contained numerous inaccuracies in citations, hyperlinks, and the application of the tribunal’s rules.
The applicant apologized, took responsibility, and undertook to refrain from utilizing generative AI when preparing additional materials.
The applicant explained that he could not retain a lawyer to assist him in this process, believed in good faith that the AI tools would provide accurate citations and reasoning, and did not mean to mislead the tribunal.
Bias motion denied
The Law Society Tribunal’s Hearing Division dismissed the applicant’s admissibility and bias motion for a lack of basis.
The tribunal found the law society’s documents admissible and relevant, considering that the rules of evidence for motions for interlocutory suspensions and variations of interlocutory suspensions were less strict than civil evidentiary rules.
The tribunal saw no bias in the panel’s review of contested evidence before determining its admissibility. The tribunal added that it did not treat the applicant in a way that led to a reasonable apprehension of bias.
The tribunal also found no bias arising from the case management hearing. The tribunal noted that the applicant mischaracterized the discussion that occurred during the hearing. Based on a transcript review, the tribunal noted that the panel’s members:
- indicated that the tribunal had an independent interest in protecting the integrity of its process
- stated that they would address the applicant’s use of AI even if he withdrew his motions raising admissibility and bias issues and seeking to vary or remove the interlocutory suspension
- had yet to make findings regarding the consequences flowing from his admission to utilizing AI and submitting materials without verifying their accuracy with sufficient care
- had yet to permit the parties to make submissions on that issue
- might consider the applicant’s AI use in applying the legal test for the motion to vary or remove the interlocutory suspension and in determining costs
The tribunal informed the applicant that it might consider the errors arising from his AI use. The tribunal gave him an opportunity to consult a lawyer and prepare for the next stage in the proceedings.
