Harwell v. WestCare Nevada, Inc.
U.S. District Court, District of Nevada · D. Nev. · Nevada bar guidance
Conduct
Pro se ADA plaintiff used generative AI to draft her complaint; the AI fabricated a nonexistent Nevada Court of Appeals case.
Consequence
Warned under Rule 11; leave to amend granted; sanctions reserved for repeat violations.
Lesson
Court will say AI-use out loud and quote ChatGPT by name when the filer admits or the order otherwise documents AI involvement.
Verified May 7, 2026
- Citation
- Harwell v. WestCare Nev., Inc., No. 2:25-cv-00689-APG (D. Nev. Mar. 6, 2026) (Gordon, C.J.)
- Decided
- March 6, 2026
Summary
Pro se plaintiff Kim Elizabeth Harwell sued WestCare Nevada, Inc. in the District of Nevada (No. 2:25-cv-00689) on an Americans with Disabilities Act claim. Chief Judge Andrew P. Gordon found that Harwell "used generative AI software to prepare her complaint" and cited a nonexistent case, "Estate of Saila v. Circle K Corp., 135 Nev. 545, 353 P.3d 541, 546 (Nev. Ct. App. 2019)," in support of a negligence theory. The court could not locate the cited authority and noted the citation numbers corresponded to unrelated cases. Gordon's order emphasizes that generative AI "often invents fake cases and legal precedent" and that "using it is no excuse to not verify the veracity of citations."
- AI tool:
- ChatGPT (R&G classified ChatGPT; court referenced generative AI generally with ChatGPT as exemplar)
What sanction did the court impose?
No monetary sanction imposed. Court warned both parties about future Rule 11 violations: "citing fake cases drafted by generative AI may violate this rule" and "failure to comply in the future may result in sanctions." On the merits, Gordon granted in part the motion to dismiss with leave to amend by April 3, 2026.
Why does Harwell v. WestCare Nevada, Inc. matter for law firms using AI?
Harwell v. WestCare is the March 6, 2026 entry in Chief Judge Andrew P. Gordon’s four-order cluster (this case, Taylor v. Las Vegas Metro, Wallace v. PennyMac, and the earlier United States v. Ponce). It is the most explicit of the four on the AI-tool question: Gordon’s order references ChatGPT by name and quotes the broader generative-AI risk that “[AI] often invents fake cases and legal precedent.”
The fabricated citation, “Estate of Saila v. Circle K Corp., 135 Nev. 545, 353 P.3d 541, 546 (Nev. Ct. App. 2019),” is a useful teaching example: it uses correct-looking Nevada Reports and P.3d page numbering and would pass a casual visual check. The defect is only visible on a Westlaw or Bloomberg lookup, which is why Gordon’s standing rule (every cited authority personally read and verified) is the operative compliance requirement, not “spot the fake.”
For firms with disability-rights or healthcare-defense practice in D. Nev., the operational takeaway mirrors the rest of the cluster: pro se plaintiffs in this district are using AI to draft complaints, and the resulting citations need a verification step on first-read regardless of how plausible they look on the page.
Implications for your firm
Operational steps a firm reading this case may wish to consider documenting. Strategic and rule-application calls belong to your firm's attorneys.
- Note the explicit ChatGPT reference in Gordon's order: when the filer admits AI use or the AI-fingerprint is unmistakable, the court is willing to name the tool. This is unusual; most AI-citation orders use 'generative AI' generically.
- Watch for AI-fabricated state-court citations using real-looking Nevada Reports and P.3d numbering. The 'Estate of Saila' fabrication used a real-format Nevada Court of Appeals citation that would pass a casual look.
- When defending an ADA case in D. Nev. before Gordon, build the citation-verification record on receipt of the complaint, not after the dismissal motion is briefed; the record is more useful if it predates the court's first warning.