Sanction teardown · Court of Federal Claims, USA · 2025-10-15
Polinski v. USA
What happened
In Court of Federal Claims, USA, 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)Plaintiff cited 'Shearson Lehman Bros., Inc. v. Wasatch Bank, 788 P.2d 514 (Utah 1990)' in his Emergency Motion for Declaratory Judgment; the court found no such case and noted the citation appears fictitious and may relate to City of Monticello v. Christensen, 788 P.2d 513 (Utah 1990).
-
Fabricated (Case Law)Plaintiff cited 'Ivey v. United States, 2019 WL 2080027 (Fed. Cl. 2019)' in his Emergency Motion to Recognize and Enforce Express Trust; the court found that citation does not lead to any actual document and treated it as a fabricated authority.
-
Fabricated (Case Law)Plaintiff cited 'Self v. Ray, 61 F.R.D. 666 (E.D. Ark. 1974)'; the court stated the citation actually corresponds to Young v. Trailwood Lakes, Inc., 61 F.R.D. 666 (E.D. Ky. 1974), indicating a misattributed or fictional citation.
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
Warning
Additional detail
"On September 3, 2025, Plaintiff filed his response to the court’s order to file copies of the cases he cited (#7). Therein, Plaintiff avers he took “concrete remedial steps” to cure the time wasted by his use of artificial-intelligence-hallucinated case citations, including “submission of the verified opinions as exhibits” (#7 at 2). Indeed, Plaintiff’s response stresses how he“obtained authentic copies” of those cases and “attached” them as exhibits. See (id.).Plaintiff did not attach any exhibits to his response to this court’s order. The court is convinced that those two case citations are AI-hallucinated. Plaintiff’s insistence that they exist—and that he provided copies of them to this court—is bewildering."
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/876/Polinski_v._USA_USA_15_October_2025.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).