Quebec courts ban generative AI from judicial reasoning and evidence assessment

Quebec's Court of Appeal, Superior Court, Court of Quebec and municipal courts jointly adopted guidelines on September 4 restricting how judges can use generative AI—and what they cannot.

Generative AI cannot replace judicial reasoning, evidence assessment, or deliberation. Judges remain solely responsible for these functions, and any AI-generated material must be independently verified. "Generative AI, however sophisticated, possesses neither judgment nor conscience and is not subject to any form of accountability," the courts said in the 10-page guidelines. "It cannot understand the human, social and legal context specific to each case."

The move came after La Presse reported in March that a court decision contained references to case law that did not exist, raising the possibility a Quebec judge had used generative AI. That incident created urgency within the judiciary to establish rules before AI tools became entrenched in the system.

Judges may use generative AI for limited administrative work: correcting, revising, rephrasing, or translating text they have already written, and locating information within specified sources. But any such use must remain under human control. "Judging cannot be reduced to a technical exercise. It is an intrinsically human function, based on deliberation, consideration of context, respect for dignity, the requirement of fairness and the exercise of moral responsibility," the courts stated.

No approved, secure institutional AI tool currently exists for judges. Existing tools rely on algorithms whose workings remain largely opaque—a particular concern given that judges carry full responsibility for their decisions' accuracy, regardless of any tool used. "No error can be attributed to the tool used. The use of a generative AI tool does not diminish the judge's obligation to verify or their full responsibility for the final result," the guidelines emphasize.