Supreme Court Criticizes B.C. Lawyer for Referring to AI-Generated Cases
A Supreme Court judge has reprimanded a British Columbia lawyer for referencing two AI-generated “hallucinations” in a legal filing and has mandated her to compensate opposing counsel for their time.
Judge Masuhara deemed it “appropriate” for Ms. Ke to remunerate opposing counsel for the time they spent uncovering that the cases she intended to cite were created by ChatGPT, a free AI system for refining conversations and automating tasks.
Although Ms. Ke retracted the AI-generated cases upon realizing their falsity, Judge Masuhara expressed concern over the incident.
“As this case has unfortunately highlighted, generative AI is not a substitute for the professional expertise expected of lawyers in the justice system,” Mr. Masuhara wrote. “Competence in selecting and utilizing technology tools, including AI-powered ones, is essential.”
Ms. Ke is the legal representative of Wei Chen, a wealthy businessman, in his divorce from Nina Zhang, residing with their three children in West Vancouver.
Related Stories
Before the ruling, Ms. Ke submitted an application for Mr. Chen’s children to travel to China, citing a case of a mother taking her “child, aged 7, to India for six weeks” and another case authorizing a “mother’s travel with her 9-year-old child to China for four weeks to visit family and friends.”
The error was discovered when Ms. Zhang’s lawyers requested copies of the cases due to their unavailability based on the citation numbers.
‘Deeply Embarrassed’
Ms. Ke issued an apology letter to an associate attending a court hearing on her behalf, but the associate failed to provide a copy to Ms. Zhang’s lawyers.
Judge Masuhara mentioned that Ms. Ke later swore an affidavit detailing her “lack of awareness” regarding the risks of using an AI program.
“I am regretful of my actions. I now understand the dangers of relying on AI-generated materials,” Ms. Ke wrote. “I recognize that I should have been more cautious and diligent in preparing the materials for this application. I extend my apologies once again to the court and opposing counsel for my mistake.”
She expressed feeling “deeply embarrassed” by her mistake, stating that the negative publicity and potential consequences made it challenging to focus, leaving her feeling overwhelmed and anxious.
Although opposing counsel requested additional costs for abuse of process, the judge denied it, believing Ms. Ke’s apology to be sincere.
“These observations do not downplay the seriousness of what occurred, which—let me be clear—I find concerning,” Judge Masuhara wrote. “However, they are pertinent to whether Ms. Ke had an intent to deceive. Given the circumstances, I conclude that she did not.”
This incident is not the first time a lawyer faced sanctions for utilizing ChatGPT.
In a similar case, a U.S. judge penalized two New York lawyers last June for including six fabricated case citations generated by ChatGPT in a legal brief.
U.S. District Judge P. Kevin Castel in Manhattan ordered lawyers Steven Schwartz, Peter LoDuca, and their firm Levidow, Levidow & Oberman to collectively pay a $5,000 fine.
The judge found the lawyers to have acted in bad faith and engaged in “deliberate avoidance and providing false and misleading information to the court.”
Reuters contributed to this report.
