Complaint against New Brunswick lawyer’s use of AI part of growing trend: experts

A New Brunswick lawyer, Denis Sawyer, is under scrutiny by the Law Society of New Brunswick for allegedly submitting a legal brief in April 2025 that contained 10 fictitious citations generated by AI. Sawyer has admitted to the oversight and accepted full responsibility, but the case is sparking broader discussions about the ethical use of artificial intelligence in the legal profession.

“

Using AI is not unethical in itself. What is unethical is using AI incompetently.

Legal experts emphasize that the issue is not the use of AI itself, but rather its incompetent application. Norman Siebrasse, a professor at the University of New Brunswick's faculty of law, noted that specialized AI tools can enhance legal practice by improving efficiency and potentially making legal advice more affordable. However, he stressed that lawyers must conduct thorough due diligence to verify the accuracy of any information provided by AI.

Amy Salyzyn, a professor at the University of Ottawa's faculty of law, echoed this sentiment, stating that the ultimate responsibility lies with the lawyer, regardless of whether they receive assistance from a human or a technological tool. The increasing frequency of errors is a growing concern, with Canadian courts and tribunals recording a sharp rise in decisions involving AI-fabricated legal materials. Courtyready, a justice watchdog, reported seven such cases in 2024, escalating to 87 in 2025, and 74 by the first half of 2026.

“

When a lawyer is submitting material to court, the buck stops with that lawyer, whether or not they receive help from a human or a technology tool.

Beyond inaccurate citations, law professors like Karen Etis from the University of Ottawa are concerned about AI's potential to undermine the reliability of legal precedents. If AI-generated cases are cited and built upon, they could create a flawed foundation for future legal arguments. Furthermore, concerns exist regarding the handling of confidential client information, as AI models often train on user data, raising privacy issues.

“

You’re destroying the reliability of precedent. Because it’s possible to set a case that someone else will then cite and someone else will then build on that and it is no longer reliable.

In response to these challenges, law schools are beginning to incorporate AI literacy into their curricula, preparing future lawyers to use these powerful tools responsibly. Siebrasse believes that proper education will empower legal professionals to leverage AI effectively while mitigating its risks.

“

These things are free because they train on your data.