Ko v Li – AI Hallucination Case Referred for Criminal Contempt Proceedings
We previously discussed the many AI hallucination cases coming up across Canada in our past blog post, What Do We Know from the Courts’ Comments on Improper Use of AI?, including the case Ko v Li, 2025 ONSC 2065. Since reporting on this case, the matter has escalated and has been referred to the attorney general for criminal contempt proceedings.
What happened? Initially, the lawyer advised the court that she had not utilized ChatGPT to prepare her client’s factum. The lawyer later advised the court that she had delegated the work to a student who had used ChatGPT. However, the lawyer has since admitted to the court that she did in fact use ChatGPT to prepare the factum. The lawyer admitted to misleading the court.
To date, there have been no appearances for the criminal contempt proceedings. The Law Society of Ontario continues to investigate the matter.
Warnings from this case:
There may be serious consequences for misleading the court, especially deliberately misleading or lying to the court.
Even if a lawyer uses AI to prepare written submissions, it is best to be forthright about its use. A lawyer should be following any AI guidelines put out by the courts or their law society, which may require disclosure of AI use.
A lawyer should not transfer blame to a student, their staff, or any other delegate. The lawyer is ultimately responsible for the work product.
The most recent decision can be read here.