- with Senior Company Executives, HR and Finance and Tax Executives
- with readers working within the Banking & Credit, Chemicals and Technology industries
In a recent decision, the Quebec Superior Court ("Court") addressed the growing issue of using artificial intelligence ("AI") tools, particularly generative AI, in court filings. The case involved Mr. Jean Laprade, a 74-year-old self-represented litigant ("Defendant"), who opposed a motion filed by Specter Aviation Limited and TPVX Aircraft Solutions Inc. ("Plaintiffs") seeking to ratify a 2021 decision rendered by the Chambre arbitrale internationale de Paris ("CAIP") involving the same parties.1
In preparing his defence, the Defendant relied on generative AI tools to draft aspects of his pleadings. However, it was found that his submissions contained references to fictitious case law and legal authorities. These irregularities caught the Court's attention, leading it to assess the issue in detail and, ultimately, to sanction the Defendant.
In this bulletin, we review the background of the case and the Court's reasoning in sanctioning the Defendant.
1. Background
The parties were involved in a dispute concerning an aircraft that had been seized in the province of Quebec since 2019. On December 21, 2021, the CAIP issued an arbitral award in favour of the Plaintiffs, which was subsequently contested by the Defendant before the Cour d'appel de Paris and then the Cour de cassation. In each appeal instance in France, the arbitral award was upheld against the Defendant.
Having obtained a firm and final arbitral award, the Plaintiffs then filed a motion before the Court to have their award recognized and ratified in Quebec. Nevertheless, the Defendant continued to contest the Plaintiffs' motion.
At the hearing, the Defendant appeared without counsel, having lost the support of his lawyers shortly before the proceedings. Acting in his own defence, he prepared and filed his submissions using, among other things, generative AI tools. Upon reviewing the Defendant's materials, the Plaintiffs discovered several irregularities, particularly regarding the cited case law and legal authorities. Many references were inaccurate, non-existent, or mischaracterized, raising immediate concerns about the reliability of the Defendant's pleadings.
When questioned about these irregularities, the Defendant admitted that he had relied on generative AI tools to prepare his legal arguments, as he no longer had professional legal support. While he expressed regret for the inaccuracies, he maintained that without the assistance of such tools, he would not have been able to prepare his case adequately.
Confronted with the Defendant's submissions containing false legal references, the Court took the opportunity to reflect on the implications of using AI tools in judicial proceedings.
2. Court's Mission to Safeguard the Legal System
In its decision, the Court reaffirmed that access to justice is both a right and a privilege. While the legal system must remain open and accessible to all, including self-represented litigants, the Court emphasized that controls and safeguards must be implemented to ensure that the justice system can effectively fulfil its societal mission. These safeguards must apply equally to all litigants, whether represented by counsel or acting on their own.
While recognizing that courts must offer some flexibility to self-represented litigants and provide them with assistance to "level the playing field", such flexibility cannot extend to tolerating false or misleading submissions. In other words, while the courts must help facilitate access to justice, the pursuit of justice must remain grounded in respect for the truth, the very foundation of public confidence in our justice system.
3. Superior Court Notice on the Use of Artificial Intelligence Tools
On October 24, 2023, the Court had issued a notice ("Notice") addressed to the legal community and the public in response to the wide-scale use of AI tools and large language models by legal professionals and the general public. The Notice outlined the key principles necessary to preserve the integrity of court submissions when using large language models (see the French version of the Notice: "Avis à la communauté juridique et public – L'intégrité des observations présentées aux tribunaux en cas d'utilisation des grands modèles de langage").
The Notice emphasized three core principles: prudence, reliability, and human oversight. The Court warned that litigants must (i) exercise caution when referring to legal sources generated through AI tools, (ii) ensure that all references to jurisprudence, statutes, or commentary come exclusively from trusted and verifiable sources such as court websites, recognized publishers, or established public databases, and (iii) subject any AI-generated content to rigorous human verification to confirm its accuracy and correctness.
4. Court Findings
In this case, the Court found that the Defendant's conduct amounted to a serious procedural breach under article 342 of the Code of Civil Procedure ("Code"). Article 342 of the Code provides that the courts may sanction any party in substantial breaches in the course of proceedings. It states:
"The court, after hearing the parties, may, on its own initiative or on an application, punish substantial breaches noted in the conduct of the proceeding by ordering a party to pay to another party, as legal costs, an amount that it considers fair and reasonable to cover the professional fees of the other party's lawyer or, if the other party is not represented by a lawyer, to compensate the other party for the time spent on the case and the work involved."
The Court held that judicial filings are "solemn" acts that demand the highest standards of diligence and accuracy whether they are submitted by a self-represented litigant or by counsel. Even if unintentionally, submitting documents containing false or fictitious case law can constitute a grave violation of the integrity of the judicial process.
While the Court recognized that the Defendant sought to defend himself to the best of his abilities with the help of AI tools, it held that he remained fully accountable for the misleading case law and legal authorities "hallucinated" by the tools he used.
Ultimately, the Defendant's conduct was found to cause unnecessary work for opposing counsel and the Court, and that his reliance on fictitious case law risked eroding public confidence in the administration of justice. The Court therefore imposed a $5,000 sanction on the Defendant, both to punish his conduct and to deter similar behaviour in the future.
5. Takeaways
Although the principles outlined in this decision are consistent with other Canadian case law, it is noteworthy that the sanction was imposed on a self-represented litigant. This case illustrates the delicate balance the justice system must maintain between ensuring that self-represented individuals have meaningful access to justice while safeguarding the integrity of the process to preserve public confidence in the judicial system.
The key takeaway is that Quebec courts are not opposed to the use of AI tools. However, litigants, whether represented by counsel or not, remain fully responsible for the content of their pleadings. They must ensure that the authorities and materials they rely on are properly verified and accurate. Based on this decision, it is clear that the Quebec courts will not hesitate to sanction those who submit false or misleading AI-generated case laws or legal authorities, regardless of whether their conduct was intentional or the result of negligence.
Ultimately, technological tools can promote access to justice, provided they are used responsibly and in accordance with the court rules of procedure. The integrity of the judicial system remains paramount, and all participants share in the duty to uphold and protect it.
Footnote
1. Specter Aviation Limited v. Laprade, 2025 QCCS 3521 (CanLII)
The foregoing provides only an overview and does not constitute legal advice. Readers are cautioned against making any decisions based on this material alone. Rather, specific legal advice should be obtained.
© McMillan LLP 2025
