What was hallucinated

Fabricated: Case Law | Cited a 2018 CRT case 'Smith v. Doe' that the tribunal states does not exist; used to argue CRT had fined participants under s.92. || Misrepresented: Legal Norm | Argued the CRT can impose fines under CRTA s.92; tribunal held the CRT has never, and cannot, impose such penalties. || Fabricated: Case Law | Referred to a non-existent Supreme Court of Canada case (1 of 3); tribunal found three such SCC citations do not exist. || Fabricated: Case Law | Referred to a non-existent Supreme Court of Canada case (2 of 3); tribunal found three such SCC citations do not exist. || Fabricated: Case Law | Referred to a non-existent Supreme Court of Canada case (3 of 3); tribunal found three such SCC citations do not exist.

Details

"Ms. Simpson referred to a non-existent CRT case to support a patently incorrect legal position. She also referred to three Supreme Court of Canada cases that do not exist. Her submissions go on to explain in detail what legal principles those non-existent cases stand for. Despite these deficiencies, the submissions are written in a convincingly legal tone. Simply put, they read like a lawyer wrote them even though the underlying legal analysis is often wrong. These are all common features of submissions generated by artificial intelligence." [...]"25. I agree with Hung Long that there are two extraordinary circumstances here that justify compensation for its time. The first is Ms. Simpson’s use of artificial intelligence. It takes little time to have a large language model create lengthy submissions with many case citations. It takes considerably more effort for the other party to wade through those submissions to determine which cases are real, and for those that are, whether they actually say what Ms. Simpson purported they did. Hung Long’s owner clearly struggled to understand Ms. Simpson’s submissions, and his legal research to try to understand them was an utter waste of his time. I reiterate my point above that Ms. Simpson’s submissions cited a non-existent case in support of a legal position that is the precise opposite of the existing law. This underscores the impact on Hung Long. How can a self-represented party respond to a seemingly convincing legal argument that is based on a case it is impossible to find?26. I am mindful that Ms. Simpson is not a lawyer and that legal research is challenging. That said, she is responsible for the information she provides the CRT. I find it manifestly unfair that the burden of Ms. Simpson’s use of artificial intelligence should fall to Hung Long’s owner, who tried his best to understand submissions that were not capable of being understood. While I accept that Ms. Simpson did not knowingly provide fake cases or misleading submissions, she was reckless about their accuracy."

Sanction teardown · B.C. Civil Resolution Tribunal, Canada · 2025-04-25

Simpson v. Hung Long Enterprises Inc.

What happened

In B.C. Civil Resolution Tribunal, Canada, a filing relied on an unnamed/unconfirmed AI tool to help draft legal argument. The court identified the following problems with the citations in that filing:

Fabricated (Case Law)

Cited a 2018 CRT case 'Smith v. Doe' that the tribunal states does not exist; used to argue CRT had fined participants under s.92.
Misrepresented (Legal Norm)

Argued the CRT can impose fines under CRTA s.92; tribunal held the CRT has never, and cannot, impose such penalties.
Fabricated (Case Law)

Referred to a non-existent Supreme Court of Canada case (1 of 3); tribunal found three such SCC citations do not exist.
Fabricated (Case Law)

Referred to a non-existent Supreme Court of Canada case (2 of 3); tribunal found three such SCC citations do not exist.
Fabricated (Case Law)

Referred to a non-existent Supreme Court of Canada case (3 of 3); tribunal found three such SCC citations do not exist.

Which AI tool

an unnamed/unconfirmed AI tool. Note: Charlotin's public database records tool attribution only where a court order, brief, or reporting on the matter states it explicitly; "unidentified" or "implied" means the record indicates AI use but does not name a specific product — we do not guess.

Outcome

Other side compensated for time spent through costs order (500 CAD) (monetary penalty: 500 CAD.)

Additional detail

"Ms. Simpson referred to a non-existent CRT case to support a patently incorrect legal position. She also referred to three Supreme Court of Canada cases that do not exist. Her submissions go on to explain in detail what legal principles those non-existent cases stand for. Despite these deficiencies, the submissions are written in a convincingly legal tone. Simply put, they read like a lawyer wrote them even though the underlying legal analysis is often wrong. These are all common features of submissions generated by artificial intelligence." [...]"25. I agree with Hung Long that there are two extraordinary circumstances here that justify compensation for its time. The first is Ms. Simpson’s use of artificial intelligence. It takes little time to have a large language model create lengthy submissions with many case citations. It takes considerably more effort for the other party to wade through those submissions to determine which cases are real, and for those that are, whether they actually say what Ms. Simpson purported they did. Hung Long’s owner clearly struggled to understand Ms. Simpson’s submissions, and his legal research to try to understand them was an utter waste of his time. I reiterate my point above that Ms. Simpson’s submissions cited a non-existent case in support of a legal position that is the precise opposite of the existing law. This underscores the impact on Hung Long. How can a self-represented party respond to a seemingly convincing legal argument that is based on a case it is impossible to find?26. I am mindful that Ms. Simpson is not a lawyer and that legal research is challenging. That said, she is responsible for the information she provides the CRT. I find it manifestly unfair that the burden of Ms. Simpson’s use of artificial intelligence should fall to Hung Long’s owner, who tried his best to understand submissions that were not capable of being understood. While I accept that Ms. Simpson did not knowingly provide fake cases or misleading submissions, she was reckless about their accuracy."

How Citation Safe would have caught this

Citation Safe runs three deterministic layers before a brief is filed: (1) does the citation exist against CourtListener's database of published opinions, (2) if quoted, does that exact language appear in the source, (3) does the cited case actually support the proposition it is cited for. Fabricated case citations fail Layer 1. Fabricated or misattributed quotations fail Layer 2 even when the underlying case is real. Misrepresented holdings — a real case cited for a proposition it does not support — are the target of Layer 3. None of these checks involve asking another language model whether the citation looks right; they are lookups and text-matches against the actual source, which is why a hallucinated citation has to survive a direct lookup against the authoritative source — not another model's opinion — to earn a VERIFIED stamp; our measured false-verify rate is published live at /quality.

Check a brief before you file it → · See our live false-verify rate

Source: https://www.damiencharlotin.com/documents/457/Simpson_v._Hung_Long_Enterprises_Inc._Canada_25_April_2025.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).