News Warner v Gilbarco AI Work Product 2026
Warner v. Gilbarco, Inc., No. 2:2024cv12333 (E.D. Mich.), is a February 10, 2026, ruling by Magistrate Judge Anthony P. Patti holding that a pro se plaintiff's materials generated using AI tools like ChatGPT are protected under the work product doctrine.[1][2][3]
Background
Pro se plaintiff Sohyon Warner brought employment discrimination claims against Gilbarco, Inc. and Vontier Corporation. During discovery, defendants sought production of all documents and information about Warner's use of third-party AI tools (including ChatGPT) for litigation preparation, including deposition details on her prompts and outputs. Warner objected, asserting work product protection under Fed. R. Civ. P. 26(b)(3)(A).[1][2]
Holdings
Work Product Protection Applies
Magistrate Judge Patti held that materials qualified as work product because a pro se litigant acts as their own "party or representative," preparing them "in anticipation of litigation or for trial" under Rule 26(b)(3)(A). No attorney involvement is required for work product protection.[1][3]
No Waiver from AI Use
The court held that waiver requires disclosure "to an adversary or in a way likely to get in an adversary's hands." Generative AI like ChatGPT is a "tool, not [a] person[], even if they may have administrators somewhere in the background." Provider access does not trigger waiver, unlike attorney-client privilege (which requires no adversary disclosure).[1][2][3]
Additional Grounds
The court also found that the discovery request was irrelevant, disproportionate under Rule 26(b)(1), and sought protected "mental impressions" and thought processes. No case law supported the defendants' theory; their argument relied on a Law360 article.[3]
Distinction from Privilege
The court carefully distinguished work product (party-focused, narrower waiver) from attorney-client privilege (third-party disclosure waives). The ruling applies specifically to work product doctrine analysis, not to privilege.[2]
Distinction from United States v. Heppner
The same day, United States v. Heppner (S.D.N.Y.) reached the opposite conclusion, holding that AI-generated documents are not protected. The split creates emerging circuit tension over whether AI platforms should be treated as third parties or neutral tools.[3]
| Issue | Warner v. Gilbarco (E.D. Mich.) | Heppner (S.D.N.Y.) |
|---|---|---|
| AI-Generated Materials | Protected as work product | Not protected |
| AI Characterization | "Tool, not a person" | Third party per terms of service |
| Key Factor | Pro se status; litigation anticipation | No counsel direction; public AI terms |
Implications
This ruling adopts a "literal, party-centric" view of Rule 26(b)(3)(A), prioritizing litigation anticipation over creation method or tool privacy risks. Litigants should document AI use as "anticipation of litigation" to invoke protection. No appeal or further docket updates have been noted as of April 13, 2026.[1][2]
References
- ↑ 1.0 1.1 1.2 1.3 1.4 Proskauer, "Michigan Federal Court Protects AI-Assisted Litigation Work Product"
- ↑ 2.0 2.1 2.2 2.3 2.4 Paul Weiss, "Federal Courts Reach Different Outcomes on AI-Generated Materials and Work Product Protection"
- ↑ 3.0 3.1 3.2 3.3 3.4 Perkins Coie, "Heppner and Gilbarco: Courts Apply Privilege and Work Product Protection to Generative AI"