Hussein v. Canada (Citizenship and Immigration)
This page is a summary and commentary on a publicly available court decision, written for general informational and educational purposes only. It is not legal advice and does not create a lawyer-client relationship with Azimi Law, regardless of how specific or applicable it may seem to your situation.
This summary may be incomplete, simplified, or not reflect the most current state of the law — the decision may be appealed, distinguished, or overtaken by later rulings. Independently verify this case against the official decision before relying on it. Nothing here guarantees the outcome of any other matter, including yours.
Do not rely on this page to make decisions about your own situation. Consult a licensed lawyer, or contact Azimi Law to discuss your situation directly.
The Facts
In an immigration judicial review, counsel for the applicants brought a motion to put new evidence before the Court and asked for more time to file the applicant's record. The motion relied on authorities the Court could not find. The Court issued several directions requiring a book of authorities. Only after the last one did counsel acknowledge using an AI research tool designed for immigration and refugee practice. The tool had produced two cases that do not exist, misstated the legal test for admitting evidence that was not before the original decision-maker, and cited a case that has nothing to do with the issue. Counsel had not checked the citations independently. Since 2024 the Federal Court has required parties to declare when generative AI has been used to prepare documents filed with it.
What the Court Decided
The Associate Judge accepted that AI is a legitimate tool but held that, in this Court, its use must be declared and its output checked by a person. The Court concluded that the use was both undeclared and hidden until after repeated questioning, which it saw as an attempt to mislead, and that this, together with the failure to verify, was a special reason to award costs on the motion. The Court also raised whether counsel should personally bear those costs and invited submissions. In a follow-up decision, 2025 FC 1138, counsel was ordered to pay a modest sum personally, with an apology and the fact that counsel was not paid for the work treated as mitigating. The decisions are about the conduct of the proceeding, not the merits of the underlying immigration claim.
This decision attracted wide attention as an early Federal Court ruling on invented AI citations in an immigration case. Because immigration judicial reviews make up a large share of the Court's caseload, it sets a clear expectation for everyone who files, whether represented or not.
Does This Affect You?
- ✓You are a self-represented litigant or a lawyer or representative who uses AI tools to help prepare Federal Court documents.
- ✓Your own lawyer or representative has used AI in your case, and you want to understand the risks.
- ✓You are a party to a Federal Court immigration matter and want to know the Court's rules on declaring AI use.
Key Takeaways
- AI use in documents filed with the Federal Court must be declared, and every citation and legal statement must be verified against real sources.
- Failing to verify AI output and concealing its use can lead to costs, including costs paid personally by counsel or a representative.
- The risk lands on the applicant too, because delays, adverse costs and credibility problems affect the case itself.