Harris v. Washington County (Purgatory Correctional Facility)

What sanction did the court impose?

Defendants' Motion for Sanctions denied as untimely (filed 49 days after the close of fact discovery). On the AI question, Magistrate Judge Kohler corrected the defense framing, noting that the court had not "admonished" Harris but rather "simply warned Plaintiff to be careful if using AI" at the November 13, 2025 hearing. The court quoted the Tenth Circuit's recent observation that "[t]here is nothing inherently problematic with the use of artificial intelligence to help prepare legal materials, but its careless use can waste both judicial resources and the opposing party's time and money, and it can damage the credibility of the legal system." The court found Harris's proposed AI safeguards "adequate to address Defendants' concerns" and said it would "address any future misuse of AI should it occur." No sanction imposed on Harris.

Why does Harris v. Washington County (Purgatory Correctional Facility) matter for law firms using AI?

Harris v. Washington County is a procedurally instructive entry in the D. Utah AI-citation caseload because the AI question is raised by the defense rather than the court sua sponte, and because the court resists rather than reinforces the defense framing. Magistrate Judge Paul Kohler’s March 24, 2026 order (ECF No. 136) denies the Washington County Defendants’ Motion for Sanctions on the threshold ground that it was filed 49 days after the close of fact discovery, well past the District of Utah’s expectations for timely discovery motions. Only after disposing of the motion on timeliness does the court turn to the AI question, and even then it does so to push back. The defense had asserted that the court “admonished Plaintiff against using fake AI citations in his filings” at the November 13, 2025 hearing; the court called that “not a completely accurate representation,” noting that the actual record showed only a warning to “be careful if using AI.” The court then quoted recent Tenth Circuit guidance that AI is not “inherently problematic” but its “careless use can waste both judicial resources and the opposing party’s time and money, and it can damage the credibility of the legal system.” Finally, the court accepted Harris’s proposed safeguards as “adequate to address Defendants’ concerns” and said it would address any future misuse “should it occur.” For partners reading the case, the lesson is that the District of Utah, like most federal districts, will treat opponent AI conduct on the same procedural footing as any other discovery sanction question: the motion must be timely, the rhetoric must match the record, and prophylactic safeguards proposed by the litigant get weight. Cross-reference: Vernieri v. Dyno Nobel Americas (D. Utah Mar. 4, 2026) (Pead, Mag. J.) (footnote-level AI hallucination warning in pro se context); Lee v. Capital One Bank USA, N.A. (D. Utah Feb. 24, 2026) (Pead, Mag. J.) (Seventh Circuit pro-se AI admonition adopted).