Lawyer Receives 6-Month Suspension For Misuse Of Artificial Intelligence (Law Society Of Ontario v. Lee)

The misuse of artificial intelligence by litigants and lawyers continues to be alarming. Although it is well-documented that generative artificial intelligence platforms have a tendency to produce inaccurate answers to legal questions, and, moreover, to produce output which references non-existing cases or which inaccurately portrays legal propositions in existing cases, these tools continue to be used recklessly.

For lawyers, the use of “hallucinated” law before a court is a serious ethical breach worthy of punishment by the bodies that regulate the licencing requirements of lawyers.

Law Society of Ontario v. Lee, 2026 ONLSTH 136 represents the first case in Canada to deal with the disciplining of a lawyer for the misuse of artificial intelligence before a court.

In this case, a senior lawyer in a family law dispute relied on a factum that had been prepared using ChatGPT. The factum contained hyperlinks to four cases, three of which did not exist. With respect to the case which did exist, the factum stated that in that decision the court had removed an estate trustee for conduct that eroded beneficiary confidence and for failing to account. This statement was wrong. An opposite finding had been made.

During the oral argument of the motion in which the factum was filed, the motions judge questioned the lawyer about whether she had used artificial intelligence to write the factum. She replied that she did not.

In response to further questioning by the judge, the lawyer further explained that she did not know if anyone in her office had used ChatGPT to prepare the factum, stating “…we do not normally use ChatGPT.” Unbeknownst to the motions judge, the responses were false.

While the motions judge ordered the lawyer to “show cause” why she not be cited in contempt of court and to attend a case conference, no contempt proceeding was held at the time because the lawyer had expressed her regret for the misuse of artificial intelligence and, after the initial oral hearing, had accepted accountability for the misuse. In a letter to the motions judge, the lawyer admitted that the factum had been prepared using AI and that she regrettably failed to independently verify the case law citations.

The lawyer explained to the motions judge that a law student had prepared the factum, and in her letter further stated:

At no point did I intend to mislead the Court or submit fictitious authorities. My reliance on these materials resulted from inadvertence and an overestimation of the reliability or emerging legal technology. I now understand and acknowledge that such tools must never substitute for personal verification by counsel, especially in documents submitted to the Court.

However, these explanations were also deceitful.

During the Law Society’s investigation, the lawyer finally admitted to the investigator that she was solely responsible for preparing the factum and using ChatGPT. The lawyer was a sole practitioner and had not employed an associate since 2020 or a law student since the summer of 2024.

Accordingly, the lawyer’s conduct raised issues about her misuse of AI and her dishonesty, both before the court and the Law Society.

The Law Society Tribunal determined that the lawyer had committed several acts of professional misconduct. She had misled the court, failed to properly serve her client and failed to act with honour and integrity.

Under Rule 3.1-2 of the Rules of Professional Conduct a lawyer is required to perform any legal services to the standard of care of a competent lawyer.

Under Rule 4.06.1 of the Rules of Civil Procedure, citations in a factum must reference the relevant paragraph, provision, or page number of the authority. A lawyer is also required to certify the authenticity of every authority cited.

The Tribunal found that the citing of non-existent authorities in the lawyer’s factum and an authority that did not support the proposition for which it was cited, breached the lawyer’s duty of providing services of a competent lawyer.

Under Rule 5.1-2(e) of the Rules of Professional Conduct, a lawyer cannot knowingly attempt to deceive a tribunal or influence the course of justice by misstating facts or law.

In this case, the Tribunal found that the lawyer knowingly deceived the court on three occasions by providing false explanations in regard to the use of AI. Over a short two-week span, the lawyer knowingly deceived the court when she replied to the motions judge that she had not used ChatGPT to prepare the factum, when she indicated that she had made the necessary inquiries of staff and only then discovered that part of the factum was drafted using AI and when she said that the factum had been prepared by a law student.

The lawyer did not disclose these incidents of dishonesty to the motions judge. They were only discovered during the Law Society’s investigation.

Under Rule 2.1-1 of the Rules of Professional Conduct, lawyers have a duty to carry on the practice of law and discharge all responsibilities to clients, tribunals, the public and other members of the profession honourably and with integrity.

The Tribunal found that the lawyer had breached this rule because she had misled the Law Society on the key issue of who had misused AI in the preparation of the factum. The lawyer only admitted to having been responsible for the drafting of the factum after the Law Society pressed her to disclose the names of staff members who may have been responsible. The lawyer had initially denied the request for information until being reminded that under section 49.3(c) of the Law Society Act that the request could not be denied.

In the result, the Law Society and the lawyer jointly submitted that her penalty should be a 6-month suspension. In accepting this penalty, the Tribunal, among other things, noted that the lawyer admitted all of the relevant facts, was a senior member of the Bar, had almost no experience with AI, and did not knowingly rely on non-existent cases when she first appeared in court. As well, no hearing had been required.

Lastly, notwithstanding the lawyer’s repeated dishonesty, the Tribunal concluded that a 6-month suspension would not bring the administration of justice into disrepute.

Based on Law Society of Ontario v. Khakhan, 2022 ONLSTH 103, the 6-month suspension was a significant suspension, particular for the lawyer, who was a sole practitioner and was nearing the end of her career.

The key takeaways from this decision are that the Law Society of Ontario will seriously punish a lawyer for the misuse of AI and that lawyers who misuse AI should accept their error immediately instead of trying to deflect blame. Although it remains unclear whether a 6-month suspension for the misuse of AI will establish a benchmark penalty for the misuse of AI, it certainly represents a potential starting point in future cases where a lawyer misleads a court as a result of citing hallucinated cases. A suspension of 6 months is generally the longest suspension that the Law Society will impose short of disbarment.

In this case, it also remains to be seen whether the lawyer will face any additional punishment as result of her dishonesty in responding to questions from the motions judge. Carriage of her contempt of court proceeding was transferred to the Attorney General of Ontario and the proceeding remains ongoing.

(This blog was not written with the assistance of artificial intelligence.) A PDF version is available to download here.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.