In Villanueva v. Las Vegas Metropolitan Police Department, No. 2:24-cv-00125-ART-DJA (D. Nev. Sept. 29, 2026), Nevada District Judge Anne R. Traum found “it was clearly erroneous or contrary to law to order the production of the entirety of {Plaintiff expert] Melton’s ChatGPT history”, granting in part and overruling in part Plaintiff’s objection by finding there were components that must be disclosed and components that were protected from disclosure.

Case Background and Judge’s Ruling

In this case involving alleged violations occurring during Plaintiff’s detention at the Clark County Detention Center, this discovery dispute arose after his expert, Tom Melton, testified that he had used ChatGPT to research and draft his report. That use had not been disclosed in response to a subpoena seeking materials considered in forming his opinions. Villanueva subsequently withheld the ChatGPT history as attorney/expert work product and trial preparation material.

Magistrate Judge Daniel Albregts ordered production of the approximately 40-page log, reasoning that Melton considered it in forming and supporting his opinions. He concluded that the log was not a protected draft because Melton had not created it and it did not itself constitute a draft of the entire report. Reviewing that decision under the “clearly erroneous or contrary to law” standard, Judge Traum found “it was clearly erroneous or contrary to law to order the production of the entirety of” the log.

Judge Traum first addressed the case summaries generated by ChatGPT. Rule 26(a)(2)(B)(ii) requires disclosure of facts or data considered by an expert in forming an opinion. That obligation extends beyond information ultimately relied upon. Because Melton consulted AI-generated summaries of excessive-force decisions in developing his opinion, those responses were discoverable. Judge Traum explained: “Because AI-generated case summaries are ‘facts or data,’ disclosure under Rule 26 requires the disclosure of all case summary responses supplied by ChatGPT.” The log also provided the defendants their only opportunity to examine allegedly hallucinated citations, quotations, and legal standards that Melton had considered.

Judge Traum reached a different conclusion regarding Melton’s questions and prompts. Comparing them to online research queries, she found that they were not themselves facts or data considered in reaching an opinion; instead, they revealed his thinking about the litigation. As she stated: “Because Melton’s questions are not ‘facts or data’ and impermissibly reveal his mental impressions and theories, the Court finds that it was clearly erroneous or contrary to law to order the production of this portion of the log.” Her analysis emphasized Rule 26’s protections for mental impressions and litigation theories.

Judge Traum similarly found that ChatGPT material used to prepare a deposition outline was protected. Melton reportedly sought an outline to anticipate questions in an excessive-force deposition. Unlike the case summaries, these responses helped him prepare to testify rather than supplying information supporting his substantive opinion. Compelling disclosure would reveal his “focus, priorities, or areas of concern to address in the deposition and at trial.” Judge Traum therefore concluded that this portion of the history was protected.

Draft report material was found to be protected as well. Melton used ChatGPT to help address subjects including the opposing expert’s qualifications and report and sometimes copied its outputs into his own report. Judge Traum emphasized that Rule 26(b)(4)(B) protects expert drafts “regardless of the form in which the draft is recorded.” She rejected the defendants’ contention that the AI-generated material fell outside that protection because it was not Melton’s own generation of a report. Judge Traum also reasoned that the material would reveal protected mental impressions even if it did not qualify as a draft.

Judge Traum conducted an in camera review but found no clear error in the magistrate judge’s decision to rely on the plaintiff’s good-faith description of the log. Addressing waiver, she wrote: “Because ChatGPT is not a person, but was an application used in this instance as a research, drafting, and editing program, the Court finds that any privileged information is not waived by Melton’s use of the program.” She added that many concerns about his methodology and reliance on ChatGPT could be addressed at trial through impeachment.

As a result, Judge Traum granted in part and overruled in part Plaintiff’s objection finding “it was clearly erroneous or contrary to law to order the production of the entirety of Melton’s ChatGPT history”, but did order some components to be disclosed.

So, what do you think? Will this ruling guide other courts on what uses of GenAI tools are protected and which are discoverable? Please share any comments you might have or if you’d like to know more about a particular topic.

Case opinion link courtesy of Minerva26, an Affinity partner of eDiscovery Today.

Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.