Sanction teardown · W.D. Washington, USA · 2026-02-04

Joann LeDoux v. Outliers, Inc.

What happened

In W.D. Washington, USA, a filing relied on ChatGPT, Claude to help draft legal argument. The court identified the following problems with the citations in that filing:

• Fabricated (Doctrinal Work)
  Plaintiff's expert reports (Kababick and Shippee) cite multiple articles that do not exist or are misattributed (court reviewed a table of ~10 such faulty citations and found errors).
• Fabricated (Legal Norm)
  Plaintiff quoted text from RCW 7.72.010(2) that does not exist and altered statutory language (Dkt. 173-1 ¶ 59); court found the quoted provision nonexistent.
• False Quotes (Exhibits & Submissions)
  Plaintiff cites a purported deposition quote from James Kababick about "post-production degradation of stimulant compounds" that does not appear in the cited transcript; court was unable to find the quote.
• False Quotes (Case Law)
  Plaintiff quoted language attributed to United States v. Hankey to support an evidentiary proposition, but the quoted phrase does not appear in Hankey.
• False Quotes (Case Law)
  Plaintiff included two quoted passages attributed to Fair v. King Cnty. that do not appear in the decision the brief cites; court found the quoted language absent.
• False Quotes (Case Law)
  Plaintiff quoted language not present in In re Washington Mut. (694 F. Supp. 2d 1192 (W.D. Wash. 2009)) and in a 2011 order (2011 WL 1158387); court found the cited phrases absent and that original use was for a different proposition.
• Fabricated (Case Law)
  Plaintiff cited a non-existent state-court decision, "Berg v. Chevrolet Motor Div., 84 Wn.2d 102, 108, 524 P.2d 226 (1974)"; court identified this citation as nonexistent and noted the correct authority was Baughn v. Honda Motor Co.
• False Quotes (Case Law)
  Plaintiff quoted language attributed to HSS Enters. v. AMCO Ins. (2008 U.S. Dist. LEXIS 31659) that does not appear in that order (the phrase "harsh sanction" was not found).

Which AI tool

ChatGPT, Claude. 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

Order to Show Cause

Additional detail

Tools disclosed later (see here).

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/1469/Ledoux_v._Outliers_USA_2026.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).