Artificial Intelligence and Discoverability
The use of Artificial Intelligence (“AI”) is continuing its wildfire like spread across almost every facet of daily life. From personal use, integration into education, and in many aspects of business operations, AI has continued to expand its influence and prevalence. While these new technological advancements can be exciting, its use in the legal field is still being explored and does not come without a unique set of challenges. Businesses must carefully consider when and how to implement AI, but compounding those decisions can be the question of how to interface AI when there is pending or current litigation. Whether a business serves the maritime, energy, transportation, construction, financial, or other sector of the economy, the issue of how AI impacts discoverability of materials looms large.
Often, industry insiders and professionals do not consider the legal privacy concerns associated with inputting materials into various types of AI systems. Data or information uploaded into an AI system may be used by third-parties for a variety of purposes. In fact, many AI platforms use data inputs to better their algorithms or to train the model. Information that is input into third-party AI platforms is given to the third-party for their use. As such, communication between a business professional and an AI system is not as private as it may appear.
Legal Considerations
Recent court decisions from around the United States have brought the issue of whether documents/information uploaded to an AI system are discoverable to the forefront of many attorneys’ minds. There is considerable uncertainty and variation in the landscape that recent federal district court decisions have created. Some courts have tended to favor enforcing protections that prevent disclosure of materials that are input into AI resources. However, other jurisdictions have indicated that certain utilizations of AI will result in waiver of disclosure protections and privileges.
- United States v. Heppner, 820 F.Supp. 3d 292 (S.D.N.Y 2026) – is considered a watershed decision regarding the discoverability of information that a client inputs into an AI resource. Heppner is a criminal decision, thus not directly applicable to civil litigation. The United States District Court for the Southern District of New York held that a criminal defendant’s voluntarily inputs into an AI resource were not protected by the attorney client privilege or the work product doctrine. The rationale employed by the district court ultimately boils down to the idea that sharing information with an AI resource is a waiver of discoverability protections because AI tools can and do share the inputs with other third-parties.
- Although only a federal district court decision, it is likely that a cascade of different rulings that follow Heppner shall either reinforce or diverge from its reasoning.
- Morgan v. V2X, Inc., No. 25–CV–01991–SKC–MDB, 2026 WL 864223 (D. Colo. Mar. 30, 2026) and Warner v. Gilbarco, Inc., 820 F. Supp. 3d 629, 632 (E.D. Mich. 2026) both discuss the issues of AI discoverability, but through the lens of whether a pro se, unrepresented by an attorney, litigant has waived disclosure protections.
- The reasoning in Warner, that ChatGPT is a tool and not a person, and therefore inputs do not constitute a waiver of the work product doctrine, provides a hopeful light that in civil litigation, a more permissive allowance of using AI tools may be allowed by trial courts.
Key Takeaways
As attorneys grow their familiarity with the nuance of AI to the practice of law, there are an increasing amount of discovery requests that demand information shared with a third-party, including open-source AI systems. Therefore, business professionals must tread carefully and are recommended to have rigorous AI training in place to avoid potential waiver of a key protection afforded to clients.
Before putting any information, documentation, or communications into an open-source model of AI, consider carefully if this information is currently, or will likely be, claimed to be privileged. Also, it is critical for businesses, and their employees, to adhere to internal AI policies that can assist in preventing inadvertent disclosure of communications or materials that are privileged. If materials have been input into an AI open-source model, it is important to share what materials have been shared, what platform the materials have been shared on, and what requests and outputs were generated in the AI tool.
It is best practice to consult with an attorney prior to inputting any details of your case into an AI system. Inputting information regarding your case into AI programs could render previously privileged information discoverable and have potentially adverse consequences to an otherwise strong position.
Disclaimer: This material is provided for informational purposes only. It is not intended to constitute legal advice, nor does it create a client-lawyer relationship between Galloway and any recipient. Recipients should consult with counsel before taking any actions based on the information contained within this material. This material may be considered attorney advertising in some jurisdictions.
