Texas Court Extends Work-Product Protection to Non-Lawyer Generative AI Chats: Implications for Enterprise AI Compliance

The Ruling in Brief

On June 3, 2026, Judge Dorfman issued a minute entry in Tate Group Automotive, LLC v. Legacy Automotive Capital, LLC, denying a defendant’s demand that the plaintiff hand over AI chat conversations. The core reasoning: under Texas rules of civil procedure, work-product protection extends to materials prepared in anticipation of litigation by or for a party — not exclusively by attorneys.

This framing matters. It means a corporate principal, a pro se litigant, or potentially any non-lawyer employee who uses a generative AI tool while preparing for litigation could, under Texas law, claim those conversations as protected work product.

The judge drew support from two 2026 federal district court rulings — one from the Eastern District of Michigan and one from the District of Colorado — both of which held that work-product protection is waived only when materials are disclosed to an adversary, or in a manner substantially likely to reach one. Feeding a prompt into a chatbot, the reasoning goes, does not automatically satisfy that threshold.

Where Texas Diverges from Federal Precedent

The ruling stands in direct contrast to United States v. Heppner, a February 2026 decision from the U.S. District Court for the Southern District of New York. In that case, the federal court held that inputting sensitive information into a third-party consumer AI platform constitutes a voluntary disclosure outside the attorney-client relationship — and therefore forfeits both attorney-client privilege and work-product protection.

Judge Dorfman explicitly disagreed. He argued that Texas procedural rules set a different standard, one that on its face extends protection to AI-assisted conversations prepared in anticipation of litigation.

This divergence is not merely academic. It signals that the legal treatment of AI-generated content will vary significantly by jurisdiction — a compliance reality that enterprise teams operating across state lines cannot afford to ignore.

The Limits the Court Still Imposed

The ruling is not a blanket shield. Judge Dorfman ordered the plaintiff to produce all discovery materials or products that had been shared with the AI tool — including materials generated under a protective order. The protection covers the conversations, not the underlying documents fed into the system.

The judge also recommended that the parties revisit their protective order to clarify whether, how, and to what extent confidential information may be shared with any AI tool or large language model. This recommendation carries a practical message: even where privilege may apply, the boundaries remain unsettled and must be actively managed.

The Public vs. Enterprise Tool Distinction Is Now Legally Consequential

Emerging case law consistently suggests that employees using public AI platforms — ChatGPT, Gemini, Claude, and similar consumer-facing tools — face a higher risk of privilege waiver than those using enterprise-grade, internally governed AI systems. The distinction is no longer just a security preference; it is becoming a legal one.

Organizations that have not yet drawn a formal line between approved enterprise AI tools and public alternatives should treat this ruling as a prompt to do so.

The Broader Landscape: Divergence Before Convergence

Courts across the United States are reaching inconsistent conclusions about how privilege doctrine applies to AI-assisted work. Texas is leaning toward flexibility. New York federal courts are leaning toward restriction. Other jurisdictions have yet to weigh in meaningfully.

This period of divergence is likely to persist for several years before appellate courts or legislative bodies establish clearer standards. In the interim, the risk exposure for enterprises is asymmetric: the cost of a poorly governed AI interaction surfacing in litigation is high, while the cost of implementing a clear AI usage policy is comparatively low.

A Closing Observation

The Tate ruling is a minute entry, not a full opinion on the merits. Its precedential weight is limited. But its directional signal is clear: courts are beginning to take seriously the idea that AI-assisted work, prepared with litigation in mind, deserves the same protective consideration as traditional attorney work product.

For enterprise AI adopters, the takeaway is not to rely on judicial protection as a compliance strategy. It is to build governance structures — clear policies, enterprise-grade tooling, updated protective orders, and litigation hold procedures — that do not depend on courts resolving questions that remain genuinely open. The law is catching up to AI usage. The organizations best positioned are those that do not wait for it to arrive.