AI and Discovery: Courts Split on Whether Use of AI Waives Protections Against Disclosure

This is the latest article from Strauss Troy concerning the use of AI as a tool in litigation.  Our last article discussed a recent ruling out of the United States District Court for the Southern District of New York in America v. Heppner, 25-cr-00503-JSR, in which the court held that documents uploaded to and created by public AI platforms were not protected from disclosure in litigation on the basis of either attorney-client privilege or the work-product doctrine.  However, as this decision was out of a single federal district court in New York, it was not binding on other courts in other jurisdictions, and different courts have espoused different opinions on whether the use of AI to review, analyze, and create legal documents waives protections against disclosure.

In Warner v. Gilbarco, Inc., No. 2:2024-cv-12333, the United States District Court for the Eastern District of Michigan recently ruled that records related to a pro se plaintiff’s use of ChatGPT to assist in preparing legal filings were protected from disclosure to the defendant by the work-product doctrine.  In this case, the plaintiff filed suit bringing discrimination and other related claims against the defendant, her former employer.  The plaintiff had acknowledged in the litigation that she used ChatGPT to help draft various filings in the case.  As a result, the defendant sought broad discovery of the plaintiff’s AI-related materials, including documents concerning her use of ChatGPT and other generative AI tools that she used to help her prepare filings in the lawsuit.  The plaintiff objected to this request, arguing that such materials were protected from disclosure.

The court agreed with the plaintiff and rejected the defendant’s request for such materials.  In so doing, the court concluded that such materials were protected from disclosure under the work-product doctrine.  The court reasoned that pro se litigants, such as the plaintiff in this case, were also entitled to invoke work-product protection in the same way represented parties are, and further concluded that the use of ChatGPT did not waive this protection because AI programs constitute “tools” rather than persons, and the work-product protections are generally only waived when disclosure of such material is made to an adversary rather than simply any third-party.

The court’s decision in Warner stands in direct contrast to the court’s decision in Heppner, and these are not the only courts that have been tasked with deciding AI-related privilege issues.  For instance, in Morgan v. V2X, Inc., No. 25-cv-01991, the United States District Court for the District of Colorado approved language in a protective order permitting AI use, so long as certain confidentiality measures were met.  In doing so, the court also held that the use of AI by a pro se litigant does not automatically waive work-product protection, distinguishing the Heppner case and aligning with the Warner court.  

These conflicting decisions make it imperative for individuals and businesses to be aware of the current state of the law governing the jurisdiction where they are located.  Depending on where one is located, the use of AI to assist in the litigation process could have different consequences, including possibly waiving protections against disclosure for certain material uploaded to or created by AI.  Our attorneys here at Strauss Troy are continuing to monitor the developments in this area of the law to stay up to date on how each jurisdiction treats the use of AI as it relates to the litigation process.  Strauss Troy’s litigators are ready to discuss with clients to ensure that they are protected and apprised of all developments in this area.

Stephen S. Schmidt: ssschmidt@strausstroy.com or 513.629.9422
Andrew D. White: adwhite@strausstroy.com or 513.629.9466