Prompts, privilege & discovery
Morgan v. V2X, Inc.
The court treated a pro se litigant’s AI-assisted preparation as work product but required him to identify the AI tool.
Prompts protected01
Case background
Material facts
Both parties used generative-AI tools to assist in litigation preparation. The plaintiff used consumer-facing AI platforms to review and analyze materials produced in discovery. A dispute arose when the defendant moved to amend the operative protective order to restrict the plaintiff’s use of AI tools on materials marked confidential and sought to compel disclosure of the specific AI platforms he used.
02
The decision
The court’s ruling
The court held that Rule 26(b)(3) protects qualifying pro se AI-assisted preparation and that a provider’s technical storage does not automatically waive protection. The plaintiff still had to name the tool because he had not shown that its identity revealed strategy. The court also amended the protective order to bar any party from uploading material designated confidential to an AI system unless the provider is contractually bound to refrain from storing inputs or using them to train or improve models, prohibit disclosure to third parties except as strictly necessary for service delivery, and permit prompt deletion of the confidential data on request.
Why it matters
What the decision means
Morgan v. V2X separates protected content from unprotected identifying information. Litigation-preparation chats qualified for protection, but the name of the AI service did not reveal strategy. Product terms and confidentiality controls remain relevant when a user uploads discovery.
Primary reading