Quebec courts issued new guidelines on September 5, 2026, declaring that generative artificial intelligence cannot be used as a substitute for a judge's own reasoning. The directive follows concerns that a sitting judge may have employed a generative AI system while drafting a recent decision. In the same statement, the courts highlighted that the province has not yet approved any secure, institutionally vetted AI tool for use by members of the judiciary.
The guidelines stress that while technology may assist with research or administrative tasks, the core analytical work of interpreting law and applying it to facts must remain the exclusive domain of the judge. The courts noted that no existing AI solution meets the security and confidentiality standards required for handling sensitive legal material. Consequently, judges are instructed to avoid relying on generative AI for any part of the reasoning that forms the basis of a judgment.
The move comes amid broader discussions about the role of AI in the legal system across Canada. Quebec's approach, as outlined in the September 5 announcement, reflects a cautious stance: the province has not authorized any institutional generative AI platform that could be deemed safe for judicial purposes. Without an approved system, judges lack a vetted resource that would guarantee data protection and integrity, reinforcing the courts' position that AI cannot replace human judgment.
The guidelines do not prohibit the use of AI for peripheral functions such as document management or language translation, provided those uses do not influence the substantive reasoning of a case. By setting clear boundaries, the Quebec courts aim to preserve the integrity of judicial decision‑making while allowing limited, non‑critical technological support. The announcement underscores the province's commitment to maintaining rigorous standards for judicial independence and confidentiality in the face of rapid AI development.