Ontario is finalizing new civil procedure rules to govern AI-generated evidence, responding to concerns that hallucinated citations and deepfakes are compromising the integrity of courtroom proceedings. The Ontario Civil Rules Committee expects to release the official text of these regulations in November 2026.
This regulatory shift directly impacts Canadian legal practitioners and litigants, as it mandates higher standards of transparency for digital evidence. LaunchLen Media views this proactive judicial-control approach as a necessary evolution to preserve the common law, though critics suggest rigid disclosure requirements could potentially increase litigation costs.
The stakes for the Canadian legal sector are significant, as the rules aim to address the growing prevalence of AI-driven errors that have already begun to appear in provincial courts — alongside wider concerns about AI-enabled fraud targeting Canadians.
The Two-Pillar Framework for Digital Evidence
The forthcoming rules from the Ontario Civil Rules Committee focus on two distinct categories of AI usage. The first pillar addresses acknowledged AI-generated evidence, where a party discloses that an AI system produced the material in whole or in part. Source: Law Times.
Under this pillar, a court may require the proponent to identify the specific AI system used, its training data, and the inputs or prompts provided. The rules prioritize identifying potential sampling bias and discriminatory bias that violates the Ontario Human Rights Code.
The second pillar addresses unacknowledged AI-generated evidence, such as deepfakes. Because fabricated audio or video can be highly persuasive, the proposed rule adopts a presumption of inadmissibility once material is identified as undisclosed AI-generated content.
Justice Peter Lauwers, who chairs the committee, stated that the premise is that judicial control is the most practical response to the rise of AI. The goal is to test the validity of AI-generated material while preventing fabricated evidence from distorting the adjudicative process.
Courts cannot prevent AI’s arrival, but they can establish appropriate safeguards to protect the integrity of adjudication and the development of the common law.
Justice Peter Lauwers, Court of Appeal for Ontario
Managing the Risks of Prompt Disclosure
The proposed rules highlight the sensitivity of prompt disclosure in litigation. Because the way a question is framed can materially bias an AI output, the committee examined how prompts influence the reliability of evidence.
Justice Lauwers noted that prompts might reveal litigation strategy and are arguably protected by solicitor-client privilege. Consequently, the rules do not compel prompt disclosure as a matter of right.
Instead, the framework leaves the issue to the discretion of the judge. If a party refuses to disclose their prompts, the court may choose not to admit the evidence if it appears distorted or unreliable.
This approach balances the need for transparency with the protection of legal strategy. It forces practitioners to consider the impact of their AI interactions on the validity of the evidence they present.
Current Challenges in Ontario Courtrooms
The legal system is already grappling with the consequences of passive reliance on AI tools. Appellate-level issues such as hallucinated citations and AI-edited quotations have become increasingly common, according to Justice Lauwers.
The committee has responded by updating the Rules of Civil Procedure to mandate the certification of citations. This ensures that lawyers are held accountable for the accuracy of the material they submit to the court.
Kathryn Manning, a litigation partner at DMG Advocates, noted that small claims court judges are reporting an increase in AI-generated materials. While some submissions are well-written, they frequently raise significant accuracy concerns.
These challenges are not limited to Ontario. British Columbia Court of Appeal Justice Susan Griffin recently warned that courts cannot provide an unlimited audience for AI-generated submissions.
The following list summarizes the key risks currently identified by the judiciary:
Passive reliance on AI leading to a loss of independent reasoning.
The emergence of hallucinated legal citations in court filings.
The use of deepfakes to create compelling but fabricated evidence.
The ‘liar’s dividend’ where authentic evidence is falsely challenged as AI-generated.
Sampling and discriminatory bias embedded in AI training data.
Analysis of the Judicial Control Approach
The Ontario approach represents a measured response to technological disruption. By focusing on judicial control rather than blanket bans, the province aims to integrate AI into the legal process safely.
Proponents argue this method preserves the flexibility of the common law, allowing judges to adapt to new technologies as they evolve. It ensures that the court remains the final arbiter of evidence quality.
Conversely, some critics might argue that the lack of bright-line rules could create uncertainty for practitioners. They may contend that the potential for judicial discretion to exclude evidence based on undisclosed prompts could increase litigation costs and create inconsistent outcomes.
Despite these concerns, the framework remains a necessary step to maintain public trust in the justice system. As AI becomes more sophisticated, the ability to test the reliability of digital evidence will be paramount.
Next Steps for Legal Practitioners
Lawyers should prepare for the implementation of these rules by reviewing their internal processes for AI usage, much as other Ontario professions are wrestling with digital reform in specialized care. Understanding the benefits and risks of these technologies is now a core component of professional competence.
Practitioners are advised to verify all legal sources against trusted databases before submission. Relying solely on AI-generated research without independent verification is increasingly risky given the new certification requirements.
The Civil Rules Committee expects to finalize the standards in the coming month. Once released, these rules will serve as a practical guide for judges and lawyers across the province.
Finally, legal teams should document their AI usage protocols to ensure they can meet potential disclosure requirements if challenged in court. Being prepared for transparency is the best way to navigate the changing evidentiary landscape.
Frequently Asked Questions
When will the new rules for AI-generated evidence be finalized?
The Ontario Civil Rules Committee expects the official text of the new rules to be released in November 2026.
Are lawyers required to disclose the prompts used to generate AI evidence?
The rules do not compel prompt disclosure as a right, but they leave the decision to the judge’s discretion. Failure to disclose could result in the evidence being ruled inadmissible if it is deemed unreliable.
What happens if a party presents unacknowledged AI-generated evidence?
The proposed rules adopt a presumption of inadmissibility for material found to be undisclosed AI-generated content, such as deepfakes.

