TL;DR:
In Villanueva v. Las Vegas Metropolitan Police Department, a September 29, 2026 order from the U.S. District Court for Nevada held that an expert’s AI-generated case summaries are discoverable, but AI prompts and deposition outlines used to develop those opinions remain protected as trial preparation or attorney work product. The decision, cataloged as ECF No. 112, distinguishes between material that informs the expert’s opinion and material that reveals the expert’s mental impressions or methodology. For trial teams, the ruling creates a practical framework for handling AI-assisted expert discovery and highlights the need for careful documentation, protocol-driven use of AI tools, and targeted production where outputs, not inputs, drive opinions. The order is part of the ongoing federal landscape on AI in litigation, with additional coverage noting a federal split on how courts treat expert prompts versus AI-generated outputs. (docs.justia.com)
What happened
On September 29, 2026, U.S. District Judge Anne R. Traum issued an order in Villanueva v. Las Vegas Metropolitan Police Department, No. 2:24-cv-00125-ART-DJA, addressing a discovery dispute over an expert’s use of a Generative AI platform to research and draft a report about excessive-force allegations. The court granted in part and denied in part objections to Magistrate Judge Albregts’ earlier rulings. The critical takeaway is that AI-generated case summaries considered by the expert in forming opinions are subject to disclosure under Rule 26, while the expert’s prompts to the AI, the deposition outline generated with AI, and AI-assisted draft material remain protected as trial preparation or attorney work product. The order explicitly categorizes the history log into three components: (1) prompts and questions posed to the AI, (2) an AI-generated deposition outline used for preparation, and (3) AI-generated outputs used to draft the report. The court found the prompts and deposition outline to be protected, but it held that AI-generated case summaries that informed the expert’s opinions must be produced. The judge’s ruling clarifies that AI-based case summaries constitute “facts or data” that informed the expert’s conclusions, whereas prompts reflect the expert’s mental impressions or strategy and are protected. The order’s date and docket details are explicit: DATED THIS 29th day of September 2026, in Villanueva v. LVMPD, case no. 2:24-cv-00125-ART-DJA. (docs.justia.com)
The Nevada decision follows an emerging federal circuit split on AI in expert discovery. A closely watched earlier Connecticut decision in Conservation Law Foundation v. Shell Oil Co. treated an expert’s AI prompts as discoverable methodology, while Nevada’s approach draws a sharper line between inputs used to form an opinion and outputs that reflect the expert’s thinking. National law firms have highlighted this divergence as shaping how litigants structure discovery plans and privilege logs when AI is part of expert work. (mayerbrown.com)
Practical implications for trial teams
AI outputs vs inputs. The Villanueva ruling confirms that AI-generated case summaries relied upon by an expert to reach opinions can be discoverable, while the prompts and other input material used to generate those outputs remain protected. Litigants should prepare to produce AI-derived substantive materials that informed the opinion, but shield the exact prompts or question-by-question inputs that reveal the expert’s thought process. This creates a clear divide between discoverable materials and protected methodology. (docs.justia.com)
Crafting a protocol for AI use in discovery. The decision underscores the importance of a disciplined, documented approach to AI usage in expert work. Companies and law firms should implement discovery protocols that (a) identify which AI-generated outputs fed into opinions, (b) preserve a separate, privilege-logged log for inputs or prompts, and (c) distinguish deposition outlines or drafting artifacts from core factual inputs. Consistent with the Nevada court’s reasoning, maintain a clean separation between outputs that can be produced and inputs that should be privileged. (docs.justia.com)
Role of 26(b)(3) and 26(b)(4)(B). The ruling rests on established Federal Rules of Civil Procedure protections for draft reports and trial preparation materials, while exercising Rule 26(a)(2)(B) and related provisions to disclose underlying “facts or data considered” by the expert. Practitioners should map AI-related materials to these categories in their privilege logs and discovery correspondence, anticipating objections and possible in-court argument over what must be produced. (docs.justia.com)
Cross-jurisdiction planning. The Nevada decision adds to a growing body of authority that is not uniformly aligned with other circuits. For litigants handling multi-jurisdictional cases or cross-venue discovery strategies, this is a reminder to tailor AI-discovery plans by jurisdiction and to monitor ongoing developments in other circuits that may impact privilege and disclosure boundaries. A contemporaneous Mayer Brown overview highlighted that a separate federal case in Connecticut had produced AI prompts as discoverable while contrasting with the Nevada approach. Litigators should track these developments for strategic planning. (mayerbrown.com)
Trial readiness and impeachment. Even where prompts remain protected, AI-generated case summaries that were relied upon by the expert can be introduced at trial through traditional channels, such as impeachment or demonstration of the methodology used to form opinions. Counsel should prepare to address potential “hallucinations” or mis-citations stemming from AI-assisted research during voir dire and witness examination, consistent with the decision’s emphasis on how outputs were used in forming the opinion. (docs.justia.com)
Training and preparation for teams. As AI tools become routine in complex litigation, trial teams should integrate AI-focused practice into their training. This includes how to question an expert about AI-aided research, how to structure cross-examination of AI-derived materials, and how to preserve the boundaries between inputs and outputs in discovery and at trial. Objection Academy’s general emphasis on objection drills and trial-readiness aligns with the broader need for disciplined handling of AI-assisted testimony in court, though practitioners should tailor any training to the jurisdiction and case-specific needs. (mayerbrown.com)
What to do now
Audit your AI workflow. If an expert used AI to research or draft opinions, create a clear separation between outputs that informed the opinion and inputs used to generate those outputs. Update privilege logs to reflect this distinction and prepare to disclose the outputs that actually fed the opinion, while seeking protection for prompts and drafting materials.
Update discovery plans. At Rule 26(f) conferences and scheduling orders, propose a structured treatment of AI materials, including specific categories for “AI-generated case summaries” (disclosable) and “AI prompts/deposition outlines/drafts” (protected). Expect pushback and be prepared to justify the protective posture based on the court’s reasoning in Villanueva.
Prepare trial strategy. Develop a plan to address AI-derived evidence at trial, including how to introduce or challenge AI-generated content and how to handle potential inconsistencies or hallucinations in AI outputs. Cross-examination questions should target the source and reliability of AI outputs and how the expert used those outputs to reach conclusions.
Consider training options. For ongoing training on objection handling and evidence issues related to AI, trial teams may look to specialized programs such as Objection Academy for drills on practical trial readiness, including how to confront AI-influenced testimony in real time, while ensuring compliance with evidentiary standards.
Objection Academy in practice
Objection Academy provides focused drills on courtroom objections, evidence handling, and trial-readiness skills. In the current AI-in-litigation landscape, robust training in how to challenge AI-assisted analyses and how to apply the Federal Rules of Evidence to AI-generated material can be a meaningful complement to the tactical adjustments spurred by Villanueva. While this timely development centers on expert discovery, the benefits of strong objection and evidence training translate across cases where AI tools influence witness testimony, reports, or exhibits.
Takeaway for litigators
The Villanueva ruling crystallizes a practical rule of the road for AI in expert discovery: preserve and protect the expert’s mental impressions and drafting work, but disclose AI-generated outputs that actually informed opinions. This split underscores the ongoing need for precise discovery log management, jurisdiction-specific strategy, and disciplined use of AI in trial preparation. As federal courts continue to refine their approach to AI in litigation, trial teams that implement clear AI-use protocols, structured privilege logs, and rigorous trial-readiness training will be best positioned to navigate evolving standards and to present AI-related materials effectively and ethically at trial. (docs.justia.com)
Sources
- Villanueva v. Las Vegas Metropolitan Police Department, No. 2:24-cv-00125-ART-DJA, Order dated September 29, 2026 (ECF 112). Justia Dockets & Filings. (docs.justia.com)
- Federal Court Splits On Disclosure Of Expert Witness AI Prompts, Mayer Brown, October 9, 2026. (mayerbrown.com)
- Conservation Law Foundation, Inc. v. Shell Oil Co. (connected to the broader AI prompts discovery discussion), Law360 coverage and public orders (May–June 2026). (law360.com)