Courts on AI / Fabricated or misquoted citations
Gragston v. Amazon LLC
United States District Court for the Southern District of Ohio · Federal district courts · OH · · 2026 WL 2184981 · No. 1:25-cv-206 · order
The Southern District of Ohio dismisses Gragston’s amended employment complaint but permits a motion for leave to amend. A footnote identifies a purported Sixth Circuit hostile-work-environment citation as AI-hallucinated and formally warns that future reliance may bring monetary sanctions or dismissal with prejudice.
- Disposition
- The court granted Amazon's dismissal motion, allowed Gragston to seek leave to amend, and formally warned him about AI-hallucinated cases.
- Recorded conduct
- Self-represented plaintiff cited a purported Sixth Circuit case with mismatched name, docket number, subject matter, and nonexistent Westlaw citation. — outcome: warning; actor: litigant in person
- Canonical record
- Incident tracker entry — verified outcome, regulator disposition and related rules
What the court wrote
Verbatim from the court's document (public domain). Ellipses mark omitted text; nothing is paraphrased.
As Amazon suggests (but does not expressly state), Gragston seems to rely on AI-hallucinated caselaw in support of this point. (See Doc. 28, #141 n.2). The Court's independent review suggests the same. Gragston cites a case that he refers to as “Latham v. Off. of Att'y Gen. of Ohio, No. 20-4089, 2021 WL 1324065 (6th Cir. Apr. 9, 2021).” (Doc. 27, #133). True, there is a Sixth Circuit case that bears that name, but it is a published decision from 2005 that is not about hostile work environment claims or the use of racial epithets in the workplace. See 395 F.3d 261 (6th Cir. 2005). And, as best the Court can tell, the only Sixth Circuit case with the docket number “20-4089” is United States v. Culver, 20-4089, 2021 WL 4258764 (6th Cir. Sep. 20, 2021), which is a criminal case about sentencing. Finally, there is no case with the Westlaw citation “2021 WL 1324065.” Accordingly, to avoid any such concerns going forward, the Court FORMALLY WARNS Gragston that any future reliance on AI-hallucinated caselaw in this litigation will result in sanctions, including monetary penalties and dismissal of this case with prejudice. ... But, in offering Gragston this opportunity, the Court reiterates its warning that any further reliance on AI-hallucinated authorities in connection with briefing in this matter will result in sanctions, including monetary penalties and potential dismissal of this case with prejudice.
Authorities cited in the passage: Latham v. Office of the Attorney General of Ohio, 395 F.3d 261 (6th Cir. 2005) · United States v. Culver, 20-4089, 2021 WL 4258764 (6th Cir. Sept. 20, 2021)