Sanction teardown · National Court of Justice at Lae, Papua New Guinea · 2025-05-09
Peter Gilmai v Abel Tol & Ors
What happened
In National Court of Justice at Lae, Papua New Guinea, a filing relied on ChatGPT to help draft legal argument. The court identified the following problems with the citations in that filing:
-
Fabricated (Case Law)Written submissions cited a non-existent case 'Application by Herman Leahy [2000] PNGLR 276'; court searches (PacLII, vLex) found no record and counsel later admitted it was generated by ChatGPT.
-
Fabricated (Case Law)Written submissions cited a non-existent case 'Kagamung v Southern Highlands Provincial Government [2001] PGNC 103'; court searches found no record and counsel admitted it originated from ChatGPT.
-
Fabricated (Case Law)Written submissions cited a non-existent case 'Peter Makeng v Peter Yama (2008) SC1012'; court searches found no record and counsel admitted it was AI-generated.
-
Fabricated (Case Law)Written submissions cited a non-existent case 'Bata Kamo v Secretary for the Department of Lands & Physical Planning (2005) N2806'; court searches found no record and counsel admitted it came from ChatGPT.
Which AI tool
ChatGPT. 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
Referral to Public Solicitor and PNG Law Society; Order to show cause
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/1018/OS_No_279_of_2024_-_Gilmai_v_Tol__Ors_250509_145925_1.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).