1. Home
  2. |Insights
  3. |Crowell Tracker of Court Rulings on Legal Privilege and Artificial Intelligence Tools

Crowell Tracker of Court Rulings on Legal Privilege and Artificial Intelligence Tools

Client Alert | 3 min read | 04.23.26

As companies and individuals increasingly embed AI tools in legal practice, courts are grappling with how to treat communications with, and information generated by, these tools. Chief among these questions is whether and in what circumstances attorney-client privilege and work-product protections as they are applied in different jurisdictions extend to AI-generated content or communications with an AI tool.

The applicability of these protections directly affects the ability to keep the information confidential later. Given the rapidly shifting legal landscape, practitioners and their clients should keep apprised of developing case law. Anyone considering using AI tools in connection with litigation, regulated activities, compliance, or other legal activities should consult counsel prior to such use.

We are actively monitoring developments in this area and will continue to update the following case tracker as new decisions become available. Please reach out to our team with questions or requests for specific advice.

 

Case Name

Case Number

Jurisdiction

Date

Stage at Time of Decision

Key Issue(s)

Relevant Holding

Sohyon Warner v. Gilbarco, Inc., Gilbarco, Inc. (d/b/a Gilbarco Veeder-Root), and Vontier Corporation

24-CV-12333

United States District Court for the Eastern District of Michigan, Southern Division

2/10/2026

Discovery-disputes stage after fact discovery had closed
  • Whether AI-use materials were protected work product; and
  • Whether using ChatGPT waived work-product protection by disclosing material to a third party.

The court denied a motion to produce documents concerning the use of ChatGPT:

  1. the AI-use materials were the pro se plaintiff’s protected work product—the court agreed that the defendant was seeking the plaintiff’s internal analysis and mental impressions subject to heightened protections afforded to opinion work product; and,
  2. for work-product waiver, the disclosure has to be to an adversary or in a way likely to get in an adversary’s hand, and ChatGPT is a tool, not a person, even if it may have administrators “somewhere in the background.” Dkt. 94 at 12.
United States of America v. Bradley Heppner

25-CR-00503

United States District Court for the Southern District of New York

2/17/2026
(following 2/10/2026 oral ruling)

Post-indictment and in the pre-trial stage (after a not-guilty plea and before a scheduled trial)
  • Whether AI-generated documents can be considered attorney-client communications; and
  • Whether AI-generated documents qualify as work product.

Privilege. The documents were not privileged for three reasons:

  1. the documents are not communications between Heppner and his counsel, because the Claude AI tool defendant used is not an attorney;
  2. the communications between Heppner and Claude were not confidential because Claude explicitly disclaims confidentiality, and users agree that the information shared with and generated by Claude can be shared with third parties; and
  3. Heppner did not communicate with Claude to obtain legal advice from Claude, and sharing documents with counsel later does not make those documents privileged.

Work Product in the Criminal Context. The documents also did not enjoy work-product protection because Heppner did not prepare the documents “at the behest of counsel,” and the AI documents did not disclose counsel’s strategy. Dkt. 27 at 12.
Morgan v. V2X, Inc.

25-CV-01991

United States District Court for the District of Colorado

3/30/2026

Discovery stage

Whether FRCP Rule 26(b)(3), which governs materials prepared in anticipation of litigation or for trial, protects a pro se party’s AI-assisted litigation work under work-product protections

Rule 26(b)(3) work-product protection can apply to a pro se party’s AI-assisted litigation work/mental impressions:

  1. Using an AI tool does not automatically waive work-product protection merely because a third-party provider stores the work product;
  2. AI interactions do not automatically compromise work-product protections, even though “AI use technically ‘discloses’ information to a third party” because “it is highly unlikely the information will fall into the hands of an adversary”; and,

Absent some showing that revealing the name of the AI tool would reveal the plaintiff’s mental impressions or case strategy, the work-product protection does not extend to the name of the AI platform plaintiff used (in contrast to the outputs), which the court directed plaintiff to disclose.

Contacts

Insights

Client Alert | 5 min read | 08.26.26

D.C. Circuit Upholds Designation of PFOA and PFOS as CERCLA Hazardous Substances

In a significant and unanimous ruling, the D.C. Circuit upheld the Environmental Protection Agency’s (“EPA”) final rule designating perfluorooctanoic acid (“PFOA”) and perfluorooctanesulfonic acid (“PFOS”) as “hazardous substances” under the Comprehensive Environmental Response, Compensation, and Liability Act (“CERCLA”), denying the petitions of seven industry groups seeking to overturn that designation.[1] CERCLA, also known as the “Superfund Statute,” allows EPA to require potentially responsible parties to fund cleanup of hazardous substances....

View All Insights