BCCA Justice Griffin: AI Filings Warning on Submission Risks

• British Columbia Court of Appeal Justice Susan Griffin issued a warning about the systemic risks posed by AI-generated court submissions, even without identifying specific errors in the case at hand.
• The ruling in 1182573 B.C. Ltd. v Gerber highlighted concerns that AI use introduces access to justice issues, requiring courts to dedicate more resources to irrelevant or hallucinated content.
• Justice Griffin noted a trend of overly lengthy, verbose, incomprehensible, and aggressively toned filings, which she suspects are often AI-generated and do not advance a party's legal case.
• Legal experts confirm these concerns are widespread across Canada, with AI use increasing judicial workload and contributing to delays in the justice system.
• Judges may not have time to extensively explain why AI-generated arguments are baseless, emphasizing the need for legal professionals to rigorously vet all submissions.

Judges will not always have the capacity to provide extensive explanations for dismissing baseless AI-derived arguments, instead needing to focus on issues that hold true legal significance in a given case.

A recent decision from the British Columbia Court of Appeal (BCCA) includes a significant BCCA Justice Griffin AI filings warning, highlighting the growing challenges artificial intelligence tools pose to the Canadian justice system. Justice Susan Griffin, in her ruling for the civil case 1182573 B.C. Ltd. v Gerber, addressed concerns stemming from a self-represented party who indicated they had relied on AI to prepare their court submissions. This development, according to Justice Griffin, introduces new AI access to justice concerns BCCA judges are increasingly facing.

Justice Griffin emphasized that judges operate with finite human resources and must allocate their time efficiently to serve the broader public interest by resolving disputes based on established legal principles and evidence. She cautioned that courts cannot dedicate limitless attention to litigants who submit AI-generated content that lacks relevance. The Justice Susan Griffin AI ruling underscored that judges will not always have the capacity to provide extensive explanations for dismissing baseless AI-derived arguments, instead needing to focus on issues that hold true legal significance in a given case.

This particular Justice Susan Griffin AI ruling is the latest in a series of Canadian court decisions that have grappled with the use of artificial intelligence by litigants and legal professionals in preparing court documents. Many prior rulings have focused on specific instances of AI-generated errors, such as hallucinated legal citations Canada, fake cases, or arguments incorrectly attributed to legitimate rulings. These cases often detailed how individuals responded after being caught submitting misleading, AI-produced filings.

However, Justice Griffin's decision takes a broader view, outlining the systemic AI-generated court submissions risks confronting the justice system, rather than identifying specific AI errors within the submissions of the parties in the 1182573 B.C. Ltd. v Gerber case. Beyond the well-documented issues of hallucinations and the tendency of some AI tools to produce lengthy lists of irrelevant case citations, Justice Griffin noted an increasing trend of courts receiving pleadings and submissions that are excessively long, verbose, and employ 'legal' language detached from established legal principles, often rendering them incomprehensible. She suspects that at least some of these materials, which do not advance a party's legal position, are AI-generated. Furthermore, she observed an increasingly aggressive tone in filings, a trend she also attributes to the expanding use of AI.

The concerns raised by the British Columbia Court of Appeal AI ruling resonate with observations from legal experts across Canada. Amy Salyzyn, an associate professor at the University of Ottawa's Faculty of Law, confirmed that Justice Griffin's comments align with sentiments she has heard from other adjudicators and judges nationwide. These concerns include the prevalence of lengthy, irrelevant, and even aggressive filings. Salyzyn noted that the use of AI is creating a substantial new burden on judicial workloads, as judges and tribunal members are obligated to receive and consider extensive materials, even if they are ultimately unhelpful.

This increased workload, coupled with potentially more combative interactions from litigants, contributes to delays and prolongs legal processes. Salyzyn indicated that while the degree may vary, this is a growing concern across Canada. Alex Mok, a lawyer at Allen / McMillan Litigation Counsel in Vancouver, echoed these worries, expressing apprehension about courts having to dedicate more resources to sifting through overly long or irrelevant submissions. He highlighted a common sentiment among lawyers in British Columbia, particularly in Vancouver, that the current system is already under strain and not functioning optimally.

Lawyers and compliance officers must rigorously vet all AI-generated content for accuracy, relevance, and conciseness before submission to Canadian courts, as this ruling signals judges will not tolerate irrelevant or baseless AI-derived arguments and may dismiss them without extensive explanation. Failure to do so risks judicial rebuke and adverse outcomes for clients.

Source: Original reporting via Canadian Lawyer