About

Nevada federal judge shields expert's ChatGPT prompts but orders AI-generated case summaries disclosed

Published
Score
15

Why it matters

On September 29, 2026, U.S. District Judge Anne R. Traum overturned a magistrate judge's order that would have required testifying expert Tom Melton to hand over his entire ChatGPT history in Villanueva v. Las Vegas Metropolitan Police Department (D. Nev., No. 2:24-cv-00125). The case is a civil rights suit under 42 U.S.C. § 1983 over the plaintiff's detention at the Clark County Detention Center. Melton, a use-of-force witness, testified at his October 1, 2025 deposition that he had used a generative AI platform to research and draft his expert report. Judge Traum granted in part and overruled in part the plaintiff's objection. She ruled that ordering production of the full history was clearly erroneous or contrary to law, but she required disclosure of certain components. Summaries Melton generated with ChatGPT and relied on in forming his opinions must be produced. His prompts, deposition-prep outlines, and draft report content remain protected.

The order does not specify which summaries Melton must produce or on what timeline. The record also does not show whether the plaintiff will seek review of the portions that were ordered disclosed. Earlier in the case, the plaintiff had not disclosed Melton's AI use in response to a subpoena for all materials the expert considered, and had withheld the history log as work product. That dispute is now resolved only in part, and the scope of compliance remains open.

Rule 26 separates "facts or data" an expert considered, which must be disclosed, from drafts and mental impressions, which are protected. The court held that this obligation reaches information the expert consulted, not just information he ultimately relied on. The case summaries were discoverable because Melton consulted them while forming his opinions. The deposition outlines were not, because they helped him prepare to testify rather than supporting his substantive opinion. Disclosing them would also reveal his "focus, priorities, or areas of concern to address in the deposition and at trial." The decision conflicts with a May 2026 Connecticut ruling in Conservation Law Foundation v. Shell Oil, where a magistrate judge ordered production of an expert's AI prompts. That ruling is widely described as the first federal decision to treat AI prompts as discoverable methodology. Nevada reached the opposite conclusion on prompts, and commentators are now describing a split. Attorneys should check whether their experts used AI tools and what those tools logged, and they should expect opposing counsel to press for prompt histories in both districts. Analysis of the ruling has appeared only in the past few days, from October 12 to 16, 2026, and courts are still setting the first working rules for expert AI use.

Sources

mail Subscribe to Law And Technology email updates

Primary sources. No fluff. Straight to your inbox.

Also on LawSnap