• Atlanta
• Boston
• Dallas
• Ft. Lauderdale
• London
• Minneapolis
• New York
• Oakland
• Philadelphia
• Washington, DC

by Jennifer Gibbs

A recent ruling from a Texas Business Court judge in Tate Group Automotive LLC v. Legacy Automotive Capital LLC, held that certain personal ChatGPT conversations qualify as protected work product under the Texas Rules of Civil Procedure, finding that the use of an AI tool such as ChatGPT does not automatically waive work product protection.

The Dispute

The case involves a 2025 dealership acquisition suit by Tate Group Automotive LLC. Third-party defendant Kris Tate used ChatGPT to assist with document review, analysis, and litigation strategy. Out of 53 ChatGPT conversations totaling approximately 1,800 pages produced in discovery, eight were withheld in their entirety on work-product grounds.

Defendants cited the 2026 federal case United States v. Heppner, in arguing that ChatGPT conversations are not work product because they were not prepared by or for counsel and/or did not reflect the strategy of counsel, and that any protections were waived by sharing information with a third-party technology vendor with no duty of confidentiality.

The Ruling

After in camera review, Judge Dorfman disagreed with the Heppner reasoning and said that he agreed with the federal cases cited by Tate — Warner v. Gilbarco and Morgan v. V2X Inc. — which noted that work product protections are waived by disclosure to an adversary or in a way that the materials would get in an adversary's hand.

Dorfman however, noted that Texas court rules have different standards for protecting attorney work product than the federal rules, adding that the state's rules "plainly appear on their face to extend that protection to Mr. Tate's chats.” Dozens of other pages not meeting the standard, however, were ordered to be produced.

Key Takeaways

• Work product protection can extend to AI-generated chats. Where prompts and outputs reflect litigation strategy or mental impressions developed in anticipation of litigation, they may qualify for protection under applicable rules.
• Using AI does not automatically waive privilege. The court analogized AI tools to other litigation-support technologies and declined to treat disclosure to an AI platform the same as disclosure to an adversary.
• Protective orders need updating. Judge Dorfman recommended the parties amend their protective order to make "unquestionably clear" how confidential information may be shared with a third-party technology vendor, such as ChatGPT.
• Confidentiality risks remain. The court left open the question of whether uploading discovery materials into ChatGPT violated the existing protective order, suggesting that litigants must carefully evaluate data-control settings and platform terms before using AI to review case materials.

Looking Ahead

Judge Dorfman acknowledged the novelty of the issue, noting that all case authorities cited by each of the parties were published in 2026, with one court calling its own ruling "a question of first impression nationwide." As AI tools become increasingly more accessible and popular, practitioners should expect continued development of the law in this area and should take proactive steps to address AI use in their protective orders and should also be prepared to discuss the use of AI with clients, vendors, and experts in every matter that might ultimately result in litigation.

The opinions expressed are those of the authors and do not necessarily reflect the views of the firm or its clients. This article is for general information purposes and is not intended to be and should not be taken as legal advice.