Accessibility Tools

Skip to main content

Two Subsequent Cases Protect Defendant’s Use of AI as Subject to Work-product Protection

Written on .

Soon after the deciding of the above-discussed case on February 17, 2026, in U.S. v. Heppner, a criminal case in the District Court for the Southern District of New York, two other courts found that such communications were nevertheless protected by Federal Rule 26(b)(3) in its work-product protection for a party’s use of AI tools for purposes of litigation.  Warner v. Gilbarco, No. 2:24-cv-12333, 2026 WL 373043 (E.D. Mich. Feb. 10, 2026) and Morgan v. V2X Inc., No. 25-cv-01991-SKC-MDB, 2026 WL 864223 (D. Colo. Mar. 30, 2026).   In the Warner case, the defendants’ argument was also rejected that any work-product protection had been waived by the plaintiff’s disclosure to ChatGPT, which the plaintiff’s argued was akin to disclosing information to an adversary.  The courts in Warner and in Morgan indicated that ChatGPT and other generative AI programs are tools, not persons.  The Morgan case did indicate, however, that the plaintiff had not shown that the work-product doctrine extended to the identification of the precise AI tool the plaintiff had used.  Thus, this information would still be subject to discovery.  Further, the Rule 26(b)(3) work product defense is limited to issues prepared “in anticipation of litigation,” and it is not clear that all such AI inquiries would come under this exemption.

Editor’s Note: The issues are new, as these three cases are the first to address whether a client employer’s communications with an AI tool are privileged.  It is likely that the Heppner case might not be binding in civil cases, as it was a criminal case that did not explicitly address work-product protection from discovery under the federal rules.  Nevertheless, some employers may want to be conservative in considering these matters, and follow the strategy below as to doing internal research on legally sensitive matters, particularly since other distinctions might conceivably be drawn in the future based on whether the AI communication was to an open or closed platform.

This article is part of our May 2026 Newsletter. 

View the newsletter online

Download the newsletter as a PDF

Get Email Updates

Receive newsletters and alerts directly in your email inbox. Sign up below.
ai, human reach out
A January 20, 2026, class action filed against Eightfold AI, Inc. in California is sending shockwaves through the employer and AI community…
danger sign, skull
A second “bombshell” affecting HR pertaining to AI is a federal court ruling in New York, that a defendant’s use of AI in researching and p…
CHAT GPT
Soon after the deciding of the above-discussed case on February 17, 2026, in U.S. v. Heppner, a criminal case in the District Court for the…
avoid, wave away
There are no easy answers to the above question, but some general observations will nevertheless be made.  First, the New York district cou…
ai visualization
Perhaps the starting point is to look at the type of AI platforms generally available.  At a recent conference about AI use for HR, speaker…
june 2026 legal immigration webinar promo graphic
The webinar will cover how to deal with a worksite enforcement action and various types of immigration enforcement activities. The webinar…