Quebec Court of Appeal Orders Man to Pay $1,000 Over ChatGPT Hallucinations He Was Warned About
Quebec's Court of Appeal has ordered a man to pay $1,000 to his ex-partner for the "loss of time" caused by fabricated legal citations his ChatGPT-drafted divorce filing contained, in a case notable for one specific detail: he'd already been warned the citations were fake before he submitted them anyway, according to Radio-Canada's reporting.
What Actually Happened in This Case
The man, representing himself without a lawyer in an asset-division dispute with his ex-partner, used ChatGPT to draft his own legal defense. Before filing his factum with the court, a lawyer informed him that the document contained AI "hallucinations," fabricated references and citations that don't correspond to real case law. He chose not to correct the text or remove the invented citations before submitting it to the court anyway.
The Court of Appeal ruled the resulting wasted time, both the court's own and the opposing side's, which had to independently verify the accuracy of citations that turned out not to exist, warranted financial compensation. The man was ordered to pay $1,000 to his ex-partner in addition to covering all legal costs associated with the matter.
Why This Is Quebec's Second Such Ruling in Under a Year
This decision isn't Quebec's first encounter with AI-fabricated court filings, but it marks a meaningful escalation in how deliberately the misconduct was treated. Quebec's first ruling of this kind came in fall 2025, when the Superior Court ordered a 74-year-old man, Jean Laprade, to pay $5,000 after his submissions in a long-running aviation dispute contained eight instances of "citations of authorities and case law that simply do not exist," a case we've referenced in our earlier coverage of AI hallucinations and why they happen.
What separates this new case from Laprade's is intent. Laprade appeared to submit fabricated citations without realizing they were fake. This latest defendant knew, was explicitly told by a lawyer before filing, and proceeded anyway, a distinction the court's ruling treats as meaningfully more serious even though the financial penalty was smaller.
Part of a Rapidly Growing Canadian Pattern
Quebec isn't operating in isolation on this issue. Similar cases have surfaced across Canada over the past year, including in Ontario and Manitoba, according to Radio-Canada's reporting, part of a pattern we've tracked closely in our own coverage, including Bill Oliver's viral incident accidentally reading AI editing instructions aloud during a New Brunswick legislative speech, and a separate New Brunswick lawyer facing a first-of-its-kind AI misuse complaint over fabricated citations in a court brief.
Quebec's Superior Court adopted a formal "AI governance framework" in September 2025 specifically to address responsible use of the technology in judicial proceedings, according to reporting from Le Devoir on the earlier Laprade decision, though that framework applies to court and legal professional conduct rather than governing self-represented litigants directly.
Why This Matters for Consumers
This case is a genuinely important cautionary example for anyone considering using ChatGPT or a similar AI tool to handle their own legal matters without a lawyer, a increasingly common choice as legal fees climb and free AI tools become widely accessible. The court's ruling makes clear that ignorance of AI's tendency to fabricate citations is treated far more leniently than proceeding after being explicitly warned.
For anyone self-representing in a legal matter, the practical takeaway is straightforward: AI tools can be a genuinely useful starting point for organizing arguments or understanding general legal concepts, but every citation, case reference, and factual claim needs independent verification against real, checkable sources before it reaches a judge. Courts are increasingly treating that verification step as a personal, non-delegable responsibility, regardless of whether a lawyer was involved in preparing the document.
Frequently Asked Questions
What happened in the Quebec ChatGPT court case?
A man representing himself in a divorce asset-division dispute used ChatGPT to draft his legal defense, was warned by a lawyer that it contained fabricated citations, and submitted it anyway. The Quebec Court of Appeal ordered him to pay $1,000 to his ex-partner plus legal costs.
Has this happened before in Quebec courts?
Yes. In fall 2025, Quebec's Superior Court fined a 74-year-old man, Jean Laprade, $5,000 in a separate case after his court filing contained eight fabricated legal citations generated by AI.
Can I use ChatGPT to help with my own legal case?
Courts have not banned AI use in legal filings, but self-represented litigants remain fully responsible for verifying every citation and fact an AI tool generates before submitting it, since fabricated legal references can result in financial penalties.
The Fast Version
Quebec's Court of Appeal ordered a self-represented man to pay $1,000 to his ex-partner after his ChatGPT-drafted divorce filing contained fabricated legal citations he had already been warned about before submitting it. The ruling follows Quebec's first such case in fall 2025, when the Superior Court fined a different man $5,000 for similarly AI-fabricated citations. Similar cases have surfaced across Canada in the past year, including in Ontario and Manitoba, as courts grapple with AI-generated misinformation in legal proceedings.
