On June 12, 2026, the Paul-André Crépeau Centre for Private and Comparative Law and the Network of Jurilinguistics Centres held the 20th Institute of Jurilinguistics at McGill University’s Faculty of Law.
The 20th Institute addressed the profound impacts of artificial intelligence (AI) on the practice of jurilinguistics. By offering powerful new tools, AI redefines translators’ and legal drafters’ professional roles. It also creates new risks for them to understand and to manage. Thus, scholars and practitioners shared their perspectives on the competent and responsible use of AI and on the training and support needed for young professionals entering the workplace.
The opening conference on the promises and pitfalls of artificial intelligence by Professor Abdi Aidid (University of Toronto) captivated the audience. Professor Aidid explained that Generative AI (GenAI) was developed without the constraints necessary to earn the trust of users. While GenAI is likely to increase productivity and to facilitate unrepresented litigants’ access to the Courts, users should know that it cannot replace human judgment. Professionals should reflect on the knowledge and skills that make their services especially valuable to society. Professor Aidid warned against certain applications of GenAI, particularly deepfakes that challenge our ability to assess the authenticity and credibility of evidence presented before the Courts.
Conferences in the first panel focused on the use of AI in jurilinguistics. Professor Marie-Hélène Girard (McGill University) presented her empirical research on the uses of AI to translate or to revise legal texts. Her results track the increasing use of GenAI by legal professionals, notably for translation. Many employers do not yet have policies in place to govern the use of AI, especially in the private sector. When presented with a translation by a legal translator and one generated by AI, a majority of legal professionals were unable to determine which was which. Moreover, they struggled to find the mistakes in the legal translation produced by GenAI.
Also in the first panel, Professor Amy Salyzyn (University of Ottawa) addressed the risks of GenAI for the development of the common law. There have been reported incidents where counsel’s or judges’ use of AI led to fabricated cases being included in court submissions and judgments. Such "hallucinated" content can corrupt citations, influence court decisions and, at the very least, waste court resources, since the validity of citations will need to be verified more carefully in the future. GenAI tools could produce more subtle influences that would pass unnoticed into the common law. Choices embedded into the tool could affect the sources that take precedence and shape users’ understanding of the law. Professor Salyzyn warned that if we do not reflect carefully on our use of GenAI, our cognitive abilities, our capacity to engage critically with the law, and the diversity of opinions and approaches to legal reasoning will be eroded.
In the first pair of parallel workshops following this panel, lawyer and librarian Katarina Daniels (Davies Ward Phillips & Vineberg) demonstrated how to use various GenAI tools tailored specifically to legal professionals. The goal is to use GenAI as a first-draft accelerator and consistency tool while preserving the intentional nuance, jurisdictional specificity, and bijural precision that define quality legal writing in Canada. In a live demonstration, participants were invited to evaluate AI-generated output critically, practise prompts and identify the moments when professional judgment must prevail.
In the parallel workshop, translator and business consultant Joachim Lépine (LION translations) offered advice on how to present to one’s client or prospective client the value of human-centred professional services in legal translation. The key is to build a relationship of trust and to emphasize the importance of communicating accurately and effectively with one’s audience in both languages. Mr Lépine suggested that translators should offer a warranty to their clients to emphasise the superior quality of their work and the added security afforded by their services.
After lunch, the second plenary panel discussed education and support for young professionals. Betty Cohen, president of OTTIAQ, discussed the training of translators in view of the major transformations driven by the popularization of GenAI tools. Ms Cohen insisted that the arrival of GenAI does not change the fundamental skillset that a translator should possess. However, it alters the workflow such that translators now intervene as “post-editors” of a translation largely generated by the tool. Ms Cohen suggested that training in translation should continue to emphasise core skills, encourage the development of subject-matter expertise and offer new training to address current practices and challenges in the use of GenAI.
Also in the second panel, Professor Catherine Déri (Université du Québec en Outaouais) reflected on the importance of educating students in the responsible, ethical and transparent use of GenAI. Universities and professional bodies have a responsibility to develop clear, comprehensive and up-to-date guidelines. Teaching activities, assignments and assessments may incorporate an appropriate use of GenAI or exclude its use, depending on the skills to be developed. Professor Déri provided examples of assignments that clarify the ethical use of AI and address the risk of plagiarism.
In the second pair of parallel workshops, jurist and translator Francie Gow (Government of Canada’s Translation Bureau) shared a method for using AI to assist in legal translation. Ms Gow insisted that the choice of tool must consider data privacy: free consumer tools should not be used to translate unpublished client documents. Ms Gow emphasized the built-in randomness of GenAI and the way it assembles a text in linear fashion, the initial word combinations leading to the next. The generated text is remarkably fluent, but it may be inaccurate, biased or beside the point. The method proposed by Ms Gow consists in feeding the tool highly relevant background materials, then prompting it to co-translate small segments at a time, while correcting the translation and providing additional prompts as the translation proceeds. This achieves meaningful speed gains without compromising legal accuracy or quality.
In the parallel workshop, Professor Aurore Benadiba (Université Laval) presented her empirical findings on legal practitioners’ use of AI. Lawyers and notaries report increased efficiency, improved accuracy, a reduction in human error, the automation of repetitive tasks, and faster data processing. Judges value the support provided for analysis, particularly in identifying trends in the caselaw and detecting unconscious biases. Professors use AI to create activities and to assist with grading.
In the closing lecture, Me France Allard (Justice Canada) shared her impressions from the day’s exchanges. She underlined the three main themes of the day: the tension between GenAI’s risks and potential, the changing role and responsibilities of jurilinguists in the face of this new technology and the need to rethink education and training. Me Allard announced the theme of the 21st Institute: the drafting and interpretation of multilingual legislation.
