Last month, we wrote to inform you of the AI-related legal developments in U.S. v. Heppner, noting that it would serve as a foundational case for deciding AI-related issues in our ever-evolving modern legal landscape. Today we bring your attention to two cases that discuss a related issue: whether a pro se party in civil litigation can be afforded work-product protection over his or her conversations with AI platforms. Both the Eastern District Court of Michigan, in Warner v. Gilbarco, Inc., and the District Court of Colorado, in Morgan v. V2X, Inc., have decided the answer to this question in the affirmative—which may surprise you. And, while at first, the Warner and Morgan decisions seem to contradict U.S v. Heppner, the three cases are easily harmonized.
As a reminder, in U.S. v. Heppner, Judge Rakoff of the Southern District Court of New York considered whether defendant Heppner’s communications with the AI-platform, Claude, in connection with a pending criminal investigation, were protected by the attorney-client privilege and/or the work-product doctrine. Where Heppner argued that he was preparing documents to defend against a potential indictment, the Court emphasized that he had prepared such documents on his own, without the direction of his counsel. And, because Claude is not an attorney, the communications did not reflect the strategy of Heppner’s attorney. Finally, because the Heppner documents were seized pursuant to a valid search warrant, the Court found that the documents were not privileged under either the attorney-client privilege or the work-product doctrine.
In Warner v. Gilbarco, Inc., the Eastern District of Michigan was tasked with deciding Gilbarco’s motion to compel disclosure of materials in a civil matter. The materials at issue were created by a pro se plaintiff using the AI platform, ChatGPT. Pursuant to Rule 26 of the Federal Rules of Civil Procedure (“FRCP”), the Court determined that these materials were protected under the FRCP. Specifically, the Court cited to FRCP 26(b)(3)(A), and recognized that, “[o]rdinarily, a party may not discover documents and tangible things that are prepared in anticipation of litigation or for trial by another party or its representative.” (emphasis in opinion). Moreover, the Court stated that “[e]ven if this information were discoverable, it is subject to protection under the work-product doctrine,” as the defendant’s motion “asks the Court to compel [p]laintiff’s internal analysis and mental impressions – i.e., her thought process – rather than any existing document or evidence….”
Similarly, in Morgan v. V2X, Inc., the Colorado District Court sought to resolve, inter alia, the question of “to what extent will work product protections apply to a pro se litigant’s use of AI.” As a preliminary matter, the Court emphasized the established principle that FRCP 26(b)(3) protects “the mental impressions, opinions, and theories of parties,” themselves, in addition to the litigation preparation materials of attorneys. It also noted that both parties—the pro se plaintiff and the represented defendant—appeared to be using AI to assist in their litigation work. Consequently, the Court expressed that “[a] reading of [the Rule] that conditions work product protection over AI materials on the involvement of counsel finds no support in the rule’s text…,” and that pro se litigants, who are held to the same standard as represented litigants, should be afforded the same protections.
In reaching its conclusion, the Morgan Court had the benefit of being able to consider both the Heppner and Warner decisions. While neither decision was binding on the Court, it determined that Morgan and Warner are distinguishable from Heppner for at least two reasons. The first being that Heppner was a criminal case, while Morgan and Warner were both civil cases governed by the FRCP. Further, pursuant to the FRCP, “the mental impressions, opinions, and theories of parties” are broadly protected. Indeed, the Heppner Court also considered the issue in the context of a search warrant, rather than discovery in a civil proceeding. The second distinction made by the Morgan Court was that “there was a gap [in Heppner] between the party and the attorney because the defendant acted entirely apart from his lawyer,” and no such gap exists for pro se litigants – who are “simultaneously the party and the advocate.” Finally, the Court noted that, while AI platforms like ChatGPT and Claude collect and store user data, “that does not eliminate all expectations of privacy or automatically waive protections.”
Like Judge Rakoff’s earlier opinion in Heppner, the Morgan Court’s explanation is comprehensive and lays the groundwork for future decisions in this space. The Court’s analysis of the three opinions, which itself even acknowledges “at first blush may appear contrary,” shows how subtle factual variations can lead to different, but still consistent, legal conclusions. Ergo, one thing is clear, plenty of questions will persist as Courts continue to grapple with the convergence of AI and the law.
Bressler, Amery & Ross will continue to keep you apprised of this and other AI-related legal developments.