Sanction teardown · SC Oklahoma, USA · 2026-05-27
State of Oklahoma ex rel. Oklahoma Bar Association v. Reeves
What happened
In SC Oklahoma, USA, 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)Cited to support that general objections are not considered; court found no case with that combination of style and proposition (fabricated use).
-
Misrepresented (Case Law)Cited as confirming broad discovery rights though the cited Federal Appendix entry did not discuss discovery; the cited authority was misapplied/misrepresented.
-
Fabricated (Case Law)Cited as refusing to delay deposition; court and plaintiff could only locate an unrelated 1939 Alabama Court of Appeals traffic decision with that style (fabricated for the proposition cited).
-
Fabricated (Case Law)Cited as rejecting an inmate's request to delay a deposition; the court found no such case or similar citation for that proposition (fabricated).
-
Fabricated (Case Law)Cited as granting a Rule 30(a)(2)(B) motion and finding no good cause to delay deposition; Westlaw number led to an unrelated maritime case and no supporting discovery authority was found (fabricated).
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
Public reprimand
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/2227/STATE_OF_OKLAHOMA_ex_rel._OBA_v._REEVES.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).