Ontario Lawyer Suspension Shows AI Citation Failures Can Trigger Professional Discipline

A six-month licence suspension in Ontario demonstrates how quickly an AI-assisted filing error can escalate into a professional-discipline matter when false authorities are followed by misleading statements to the court and dishonesty with the regulator.

On July 16, 2026, the Law Society Tribunal found that Toronto lawyer Mary Hyun-Sook Lee committed professional misconduct after submitting a factum prepared with generative AI that contained nonexistent or irrelevant case law. The Tribunal ordered a six-month suspension and C$10,000 in costs.

The panel did not treat this as a simple verification failure. It identified three distinct forms of misconduct:

• Failure to serve the client competently (Rule 3.1-2)
• Deliberately misleading the court (Rule 5.1-2)
• Dishonesty with the Law Society about the use of AI (Rule 2.1-1)

Why the distinction matters

Many AI citation cases focus narrowly on whether the lawyer verified the authorities. The Ontario decision adds an important layer: a lawyer’s response after discovering an AI-related error can itself become independent grounds for discipline.

The panel treated the lawyer’s statements to the court and to the regulator as separate from the initial citation errors. That framing has direct implications for how firms design escalation and disclosure protocols when AI tools produce problematic output.

Practical implications

The decision supports several concrete controls:

• Verify the filed version, not just earlier drafts.
• Preserve the drafting record so the firm can determine what the AI produced and what was changed.
• Define an escalation path that includes timely correction and accurate disclosure to the court or regulator.
• Treat regulatory responses with the same care as representations to a court — the dishonesty finding shows these statements can become part of the misconduct case.

Bottom line

AI citation problems are no longer limited to warnings or monetary sanctions. They are now reaching professional discipline, with licence suspensions based on a combination of AI-generated authorities, misleading advocacy, and lack of candour during the regulatory process.

Verification remains essential, but it is only the first control. Legal teams also need reliable records, a clear correction process, and a plan for accurate responses when a court or regulator starts asking questions.

This article summarizes a published professional-discipline order. It does not constitute legal advice.

Sources

• Law Society Tribunal order (July 16, 2026)
• Tribunal case-document guidance
• Clearon AI Courts and AI tracker