Legal briefings

Case law, analysis, and governance references

The first federal decisions on AI privilege establish a clear pattern: process design, attorney direction, and confidentiality posture determine whether AI-assisted work is protected. This page tracks the primary sources that inform Saidebar architecture.

Executive summary

Privilege is process-dependent

Courts evaluate who directed the AI use, for what purpose, and whether confidentiality was preserved. Attorney involvement at the front end is the key differentiator.

Workflow controls matter more than model choice

Enterprise buyers need supervision, review, and governance built into the operating path. Model performance alone is not a defensible position.

Policy pressure is increasing

Multiple states are evaluating restrictions on AI legal and medical advice. Governance-ready systems become a procurement and regulatory advantage.

Latest updates

As of this week, we have not identified a new published federal opinion in the last seven days that squarely changes the AI + discovery + privilege landscape.

The leading authority remains United States v. Heppner (S.D.N.Y. Feb. 17, 2026), with Reuters identifying the opinion as 2026 WL 436479 in its legal analysis coverage.

The key counterpoint remains Warner v. Gilbarco (E.D. Mich. Feb. 10, 2026), where work-product protection was preserved on those facts. The practical takeaway is unchanged: privilege outcomes are process-dependent.

Primary case materials

Federal decisions that directly address AI privilege, work product, and discovery.

Privilege

United States v. Heppner

S.D.N.Y., Feb 17, 2026

A represented defendant used Claude AI to draft defense strategies without attorney direction. The court held that privilege did not attach because no attorney was involved at the front end, the AI platform is not an attorney, and the platform’s privacy policy permitted data disclosure. Reuters coverage cites the opinion as 2026 WL 436479. The court noted that attorney-directed AI use may have preserved a privilege claim.

Takeaway: Consumer AI use outside counsel direction can undercut both privilege and work product arguments. Process and direction matter.

View document
Work product

Warner v. Gilbarco

E.D. Mich., Feb 10, 2026

A pro se plaintiff used AI tools in litigation preparation. The court protected the materials as work product, rejecting the argument that AI platform use constituted waiver. The court emphasized that the plaintiff was the party, and her internal analysis and mental impressions were protected.

Takeaway: Work product outcomes are fact-specific. Process design, party status, and the nature of materials drive discovery results.

View document
Discovery

Tremblay v. OpenAI

N.D. Cal., June 24, 2024

In a discovery dispute, the court compelled production of AI prompts, outputs, account settings, and testing process. The court held that placing a subset of test results in the complaint waived work product protection over related negative results.

Takeaway: AI prompts, outputs, and testing process can become central evidence in litigation. Account settings and interaction history are discoverable.

View document

Analysis and market context

Secondary sources explaining why attorney-supervised AI workflows are accelerating.

Reuters: AI tools as third-party disclosure risk

Recent legal analysis of United States v. Heppner emphasizing that uploads to public AI platforms may be treated as third-party disclosure, undermining confidentiality and privilege.

Read analysis

JD Supra: A Tale of Two Cases

Comparative analysis of Heppner and Warner, framing practical implications for corporate counsel and best practices including employee policies, litigation hold updates, and closed AI tool requirements.

Read analysis

Bond, Schoeneck & King: No Counsel, No Privilege

April 2026 practitioner alert framing Heppner as a warning that unsupervised client-side AI use may fall outside attorney-client privilege.

Read analysis

Blakes: AI and legal privilege practical considerations

April 2026 practical guidance for in-house teams and firms on preserving confidentiality and privilege posture in AI-assisted legal workflows.

Read analysis

Missouri Lawyers Media: Heppner discovery risk summary

Concise April 2026 coverage of Heppner focused on litigation and discovery exposure when AI interactions are outside attorney-directed channels.

Read analysis

Standards and governance

Professional duty and risk management frameworks that inform our operational approach.

Fed. R. Civ. P. 26

Federal discovery rules and work product doctrine framework.

View reference

ABA Model Rule 1.6

Confidentiality of information and duty to prevent unauthorized disclosure.

View reference

NIST AI Risk Management Framework

Federal framework for trustworthy AI controls, risk management, and governance.

View reference

Policy watchlist

We are monitoring legislative proposals across multiple states, including New York and Maryland, where stakeholders are advocating restrictions on AI legal and medical guidance.

Our position: preserve responsible public access to research tools while requiring stronger safety engineering, transparent disclosure, and professional supervision for higher-risk usage. As specific bill numbers and committee materials become available, they will be tracked here.

Brief your team

We provide custom briefings for law firm leadership, legal operations, compliance, and procurement.