Court Rules Expert’s AI Prompts Are Fair Game Under Rule 26

A federal court yesterday ordered a plaintiff to produce the AI prompts its expert used to analyze documents in a pending environmental case, holding that the prompts are part of the expert’s methodology and fully discoverable under Rule 26. The ruling is among the first to directly address whether expert AI prompts are discoverable, and it adds an important new piece to an emerging framework that courts have been constructing to address the use of AI in litigation.

Background

Conservation Law Foundation v. Shell Oil Co. is an environmental enforcement action pending in Connecticut. Civil No. 3:21-cv-00933 (VDO) (D. Conn.). In preparing her expert report, CLF’s expert Dr. Naomi Oreskes used AI tools to review Shell’s document production and identify a subset of documents for closer analysis. When Shell and its co-defendants sought discovery of the prompts Dr. Oreskes used in that process, CLF resisted on three grounds. Magistrate Judge Thomas Farrish rejected all three.

The Court’s Analysis

CLF first argued that AI prompts fall entirely outside the scope of Rule 26(b) discovery. The court disagreed. Expert methodology is fair game for discovery, and the process by which Dr. Oreskes used AI to narrow defendants’ document production into a workable subset was part of that methodology and therefore discoverable.

CLF next argued that a Rule 29 discovery agreement between the parties protected the prompts from disclosure. The agreement excluded discovery of “expert notes, drafts, or communications needed by, and made during, the report drafting process,” and CLF characterized the AI prompts as “notes.” The court declined to read the agreement that broadly. Before a court will deny otherwise-relevant discovery based on a Rule 29 agreement, that agreement “must be quite clear,” and that court concluded that characterizing AI prompts as “notes” did not meet that standard.

Third, CLF argued it had nothing additional to produce because Dr. Oreskes had not used “prompts,” only “search terms,” which CLF said had already been produced. The court noted that a party’s good-faith representation that no additional responsive materials exist ordinarily resolves the matter. But that principle gives way when the requesting party has an evidence-backed reason to doubt the representation. Here, an earlier declaration by Dr. Oreskes’ research assistant had specifically referenced “prompt[s],” giving defendants an evidence-backed reason to doubt CLF’s representation that no additional responsive materials existed.

Where This Fits

Today’s ruling is the latest in a line of decisions addressing AI prompts in litigation. Courts have now addressed attorney-crafted prompts (likely protected as opinion work product), prompts generated without counsel’s involvement (protection depends in part on whether the platform maintains confidentiality of user inputs), expert prompts as part of disclosed methodology (today’s case), and the use of AI tools on discovery materials generally (where courts are requiring closed AI tools through protective orders). The thread running through all of these decisions is that protection turns less on the technology and more on who created the prompts, for what purpose, and under what terms of service. We address the full framework in a companion post.

What This Means in Practice

The decision has practical implications for both sides of an expert dispute. If you are retaining an expert who will use AI, treat the prompts as part of the expert’s methodology from the beginning. Preserve them, understand how they are used, and assume they may need to be disclosed. Prompts that reflect a disciplined analytical approach may strengthen the credibility of the expert’s work; poorly defined prompts may invite challenges to the methodology itself.

Address AI use explicitly in engagement letters, discovery protocols, and Rule 29 agreements. The parties here had an agreement that covered some of this ground, but its silence on AI tools specifically was enough to defeat CLF’s position. An agreement that does not expressly address AI prompts and outputs may not protect them from discovery.

If you are opposing an AI-assisted expert, the prompts are a legitimate discovery target. They can reveal the assumptions underlying the analysis, the parameters the expert set for document selection, and any disconnect between what the AI was instructed to do and what the expert ultimately claims the documents show.

The order in Conservation Law Foundation v. Shell Oil Co. is available on PACER at ECF No. 970, D. Conn. May 18, 2026.

© Arnold & Porter Kaye Scholer LLP 2026 All Rights Reserved. This Blog post is intended to be a general summary of the law and does not constitute legal advice. You should consult with counsel to determine applicable legal requirements in a specific fact situation.