In re: Sanctions Order of Kenney
Court of Appeal of Louisiana, Fifth Circuit · La. App. 5 Cir. · Louisiana bar guidance
Verified September 29, 2026
- Citation
- In re Sanctions Order of Kenney, No. 25-C-389, 2025 WL 2986582 (La. App. 5 Cir. Oct. 23, 2025)
- Decided
- October 23, 2025
Summary
Plaintiff's counsel Connie P. Trieu, in a medical-records invasion-of-privacy suit, filed pleadings citing three entirely fabricated Louisiana cases (Burns v. State, Smith v. Christus St. Patrick Hosp., and Doe v. Southwest Louisiana Hosp. Ass'n), misrepresented the holdings of two genuine cases, and included a false quotation. Trieu stated that her paralegals and law clerks had cross-checked the citations using Google, ChatGPT, and Microsoft Copilot and printed AI-generated case summaries, which she said she read as cases. After defense counsel flagged the errors, Trieu's supplemental memorandum and "errata" cited a genuine case (Thomas v. Regional Health System of Acadiana) with a wrong reporter citation and a misstated holding, which the court called "another apparently AI-generated fabrication." The trial court (Judge Nancy A. Miller, 24th JDC, Jefferson Parish) sanctioned Trieu under La. C.C.P. art. 863, and the Fifth Circuit (Judge Timothy S. Marcel writing) affirmed.
- AI tool:
- ChatGPT, Microsoft Copilot, and Google
- Sanction amount:
- $1,368
What sanction did the court impose?
$1,368 in attorney's fees plus the defendants' court costs of the sanctions motion (trial court), the costs of the writ proceeding, 3 hours of CLE training on the ethical use of generative AI (certificate due to trial court by December 31, 2025), and referral to the Office of Disciplinary Counsel for the Louisiana Attorney Disciplinary Board for evaluation of potential Rules of Professional Conduct violations.
Why does In re: Sanctions Order of Kenney matter for law firms using AI?
The Fifth Circuit said that sanctions for AI-generated fake citations are an issue that “is res nova for this Court and, to our knowledge, has not been considered by any other Louisiana appellate court.” The attorney told the court her staff “cross-checked the citations with three different sources, Google, ChatGPT, and Microsoft Copilot,” the same kind of tools that produced them. For managing partners, the operative warning is in the court’s discussion of the AI-generated case summaries: the printouts carried the disclaimer “AI responses may include mistakes,” omitted the names of the judges, and used “lists and bullet points which rarely appear in the formatting of reported cases,” yet counsel read them as cases. The appellate court added its own measures to the trial court’s fee award (CLE on the ethical use of generative AI and a report to the Office of Disciplinary Counsel), and it declined damages for a frivolous writ “[i]n this single instance” because the matter “presents a res nova legal question,” while warning that “[f]uture appeals or writs that arise from similar cases involving AI-generated fake citations may be deemed frivolous and subject to additional penalties depending on the facts and circumstances presented” under La. C.C.P. art. 2164.