Warner v Gilbarco
Warner v. Gilbarco, Inc. (Case No. 2:24-CV-12333, E.D. Mich.) is an employment discrimination case in which a federal magistrate judge ruled that a pro se plaintiff's AI-assisted litigation materials are protected under the work product doctrine. Issued on the same day as United States v Heppner — which reached the opposite conclusion — the paired rulings represent the first judicial determinations on whether AI-generated materials qualify for legal privilege protections.[1]
Parties
Plaintiff
- Warner — pro se litigant (acting as her own counsel) in an employment discrimination action
Defendants
- Gilbarco, Inc. and others
Court
- Court: United States District Court for the Eastern District of Michigan
- Case No.: 2:24-CV-12333
- Citation: Warner v. Gilbarco, Inc., 2026 WL 373043 (E.D. Mich. Feb. 10, 2026)
Factual Background
Warner, a pro se employment discrimination plaintiff, used generative AI tools including ChatGPT to prepare litigation materials in connection with her lawsuit against Gilbarco. Defendants moved to compel production of "all documents and information concerning her use of third-party AI tools in connection with this lawsuit," arguing that using public AI tools waived attorney-client privilege and work product protection because the AI interactions constituted disclosures to third parties.[2][3]
Ruling (February 10, 2026)
A federal magistrate judge denied the motion to compel, holding that the AI-assisted materials were protected under the work product doctrine.[2][1]
Work Product Doctrine — Protected
The court held that Warner's AI-assisted materials qualified as work product under Federal Rule of Civil Procedure 26(b)(3)(A) because they were prepared in anticipation of litigation. As a pro se litigant acting as her own counsel, Warner could assert work product protection over her litigation preparations regardless of the medium used.[2][4]
Key reasoning:
- AI tools are tools, not persons — The court rejected the argument that using ChatGPT constitutes disclosure to a third party, stating that "AI tools are tools, not persons"[2][1]
- Work product protects thought processes regardless of medium — The doctrine protects a party's litigation-related thought processes and preparation, regardless of whether a pen, computer, or AI tool was used[2]
- No waiver from AI provider access — Work product waiver requires disclosure "to an adversary or in a way likely to get in an adversary's hand." Using a public AI tool does not meet this standard.[5][6]
Attorney-Client Privilege — Not Extended
The court did not extend attorney-client privilege to the AI-assisted materials, noting that the privilege applies to human attorney-client communications. However, this was not dispositive because the work product doctrine provided independent protection.[3][1]
Comparison with United States v. Heppner
On the same day (February 10, 2026), Judge Rakoff in United States v Heppner (S.D.N.Y.) reached the opposite conclusion, denying both privilege and work product protection for AI-generated materials. Key distinctions:[1][7]
| Factor | Warner (E.D. Mich.) | Heppner (S.D.N.Y.) |
|---|---|---|
| Case type | Civil employment | Criminal fraud |
| Counsel status | Pro se (own counsel) | Had counsel, acted independently |
| Work product | Protected — pro se can assert | Denied — no counsel direction |
| Privilege | Not extended (not needed) | Denied — no attorney-client communication |
| AI characterization | "Tools, not persons" | Third-party disclosure |
The critical distinction is Warner's pro se status: as her own counsel, her litigation preparations qualify as work product even if made with AI assistance. Heppner, who had counsel but acted independently of counsel, could not establish the necessary connection to attorney-directed work.
Significance
Warner v. Gilbarco establishes that:
- Pro se litigants' AI-assisted materials can qualify for work product protection[2]
- Courts may treat AI tools as tools rather than third-party recipients for waiver analysis[1]
- The medium of preparation (AI vs. manual) does not determine work product protection — the purpose and context of creation matter[4]
- Work product and attorney-client privilege are distinct doctrines that must be analyzed separately for AI-generated materials[6]
Case Information
- Court: U.S. District Court for the Eastern District of Michigan
- Case No.: 2:24-CV-12333
- Citation: 2026 WL 373043 (E.D. Mich. Feb. 10, 2026)
- Ruling Date: February 10, 2026
See Also
- United States v Heppner — Paired ruling denying AI-generated materials privilege
- News Article on Warner v. Gilbarco
- Cases — List of AI-related litigation
References
- ↑ 1.0 1.1 1.2 1.3 1.4 1.5 Perkins Coie, "Heppner and Gilbarco: Courts Apply Privilege and Work Product Protection to Generative AI," February 2026
- ↑ 2.0 2.1 2.2 2.3 2.4 2.5 Proskauer, "Michigan Federal Court Protects AI-Assisted Litigation Work Product," February 2026
- ↑ 3.0 3.1 JD Supra, "Two Courts, Two Answers: When Does Using Generative AI Waive Privilege?," February 2026
- ↑ 4.0 4.1 Kirkland, "Two Federal Courts Chart Diverging Paths on the Discoverability of LLM Interactions," March 2026
- ↑ Cite error: Invalid
<ref>tag; no text was provided for refs namedprokinscoie - ↑ 6.0 6.1 Gray Reed, "Differing Federal Court Rulings on AI-Generated Documents," March 2026
- ↑ Paul Weiss, "Federal Courts Reach Different Outcomes on AI-Generated Materials," March 2026