A federal court in Nevada ruled that an expert witness’s complete ChatGPT history is not automatically discoverable in litigation, drawing a distinction between AI-generated research relied upon in forming opinions and materials that reveal an expert’s mental impressions or trial strategy. In Villanueva v. Las Vegas Metropolitan Police Department, the court held that AI-generated case summaries used by the expert to develop opinions must be disclosed because they constitute “facts or data” considered in the expert’s analysis. However, the court protected the expert’s prompts, questions, deposition-preparation materials, and AI-assisted report drafts, finding that these materials reflect protected work product and trial preparation.

The decision provides important guidance for attorneys, experts, and organizations using generative AI in litigation. The court emphasized that discoverability depends on the purpose and substance of the AI-generated material rather than the technology itself. The ruling encourages legal teams to establish clear protocols governing AI use, preserve relevant research outputs, validate AI-generated information, and distinguish discoverable factual materials from protected strategy and draft work. The court also noted that while certain AI-generated content may be protected from disclosure, attorneys remain responsible for verifying the accuracy and reliability of AI-assisted work.