Quebec’s courts have issued new guidelines restricting how judges can use generative artificial intelligence, emphasizing that AI may assist with limited tasks but cannot replace human judicial reasoning, evidence assessment or decision-making.
The guidelines were jointly adopted by the Quebec Court of Appeal, Superior Court of Quebec, Court of Quebec and municipal courts. They state that judging must remain the responsibility of human judges, who are accountable for the reasoning and conclusions contained in their decisions.
Under the new rules, judges may use generative AI for limited administrative and supporting tasks, including correcting, rephrasing or translating text they have already prepared and locating information within specified sources. Any AI-generated material must be independently verified by the judge.
However, AI cannot be used to conduct substantive legal reasoning, assess evidence or credibility, determine the outcome of a case or formulate the reasoning behind a judgment.
The courts’ position follows concerns raised earlier this year after a Quebec court decision reportedly contained references to nonexistent case law, raising questions about whether generative AI had been used in preparing the decision.
Quebec’s judiciary also highlighted concerns about confidentiality and data security. Judges are warned not to enter draft judgments, case notes or other legally protected information into unauthorized commercial AI systems.
The development reflects a broader debate across Canada about how courts should regulate artificial intelligence. Canadian courts have increasingly developed policies requiring caution, verification and meaningful human oversight when AI tools are used in legal work.
Quebec’s Superior Court has also experimented with AI in a controlled pilot program, allowing participating judges to use Microsoft AI applications for tasks such as writing assistance, translation, legislative research and legal citations—but not for drafting rulings or performing judicial decision-making.
The new guidelines therefore do not amount to a complete ban on AI in Quebec’s courts. Instead, they establish a clear boundary: AI can support judges, but it cannot become the judge.
The policy underscores the judiciary’s emphasis on human judgment, fairness, accountability, confidentiality and judicial independence as artificial intelligence becomes increasingly common in the legal system.
Swifteradio.com
