⚡ CASE DIGEST
Zhang v. Chen — BC Supreme Court, 2024 BCSC 285
A BC Supreme Court judge discovered that two cases cited by a lawyer in a family law parenting application did not exist — they were generated by ChatGPT. The judge ordered the lawyer to pay costs personally (not the client) and directed a review of all her other pending files — one of the earliest Canadian decisions to impose personal cost consequences directly on an AI-hallucinating lawyer.
Category: Hallucinations & Sanctions | Jurisdiction: Canada (British Columbia) | Read time: 5 min
① Case at a Glance
| Full Citation | Zhang v. Chen, 2024 BCSC 285 |
| Court | Supreme Court of British Columbia |
| Date | February 2024 |
| Category | Hallucinations & Sanctions |
| Jurisdiction | Canada (British Columbia) |
| AI Tool Used | ChatGPT |
| Underlying Dispute | Family law — parenting time / children’s travel |
| Source | CanLII — 2024 BCSC 285 |
② Background
Zhang v. Chen was a family law proceeding in the BC Supreme Court involving, among other issues, whether the children of the parties could travel to China with one parent. The applicant’s lawyer filed written submissions that included case citations to support the legal arguments. During the hearing, the court or opposing counsel attempted to locate two of the cited authorities — they could not be found on CanLII, Westlaw, or any other legal database. The lawyer acknowledged she had used ChatGPT and that the two cases had been generated by the AI tool — they did not exist.
③ The AI Issue
Should the financial burden of AI-hallucinated citations fall on the client (who had no knowledge of the AI research method) or on the lawyer personally? And what institutional steps should the court take to prevent the same lawyer from causing similar harm in other files?
④ What the Court Decided
- Personal costs ordered against the lawyer — not the client [wasted costs / solicitor-own-client costs order, aimed at lawyer misconduct rather than client liability].
- The court directed a review of the lawyer’s other pending files to assess whether ChatGPT-generated fictional citations appeared elsewhere — a broad institutional response to systemic AI misuse risk.
- Special costs (the highest BC category, reserved for reprehensible conduct) were declined — the court found the conduct was negligent rather than deliberate.
- The decision confirmed that in Canadian family law — where children’s welfare is paramount — AI-generated errors attract direct personal consequences for counsel.
⑤ Key Quote
“Counsel’s reliance on ChatGPT-generated authorities without verification is not a burden that the opposing party — or the court’s time — should bear. Costs will be payable personally by counsel.”
— BC Supreme Court, Zhang v. Chen, 2024 BCSC 285
⑥ The India Angle
Zhang v. Chen is particularly instructive for Indian family law advocates. Family courts in India under the Family Courts Act, 1984 handle matters involving children’s welfare, maintenance, custody, and adoption — proceedings where the court acts as parens patriae and the child’s interests are paramount. Indian courts have inherent power under Section 151 CPC to order costs de bonis propriis (personally against the advocate) when the advocate’s conduct causes unnecessary expense — functionally identical to the BC personal costs order. A client whose family law case is prejudiced by a lawyer’s careless use of AI research tools has a viable complaint under Section 35 of the Advocates Act, 1961. Gross negligence in research — especially in matters affecting children — would be treated as “professional misconduct” under the BCI disciplinary framework.
⑦ Quick Takeaways
- Personal costs — not client costs — fall on the lawyer when AI hallucinations waste court time.
- Courts can trigger systemic file reviews of a lawyer’s entire practice after discovering AI misuse.
- Family law is especially high-risk for AI errors: children’s welfare cases demand verified, accurate law.