Morgan v. V2X, Inc., No. 25-cv-01991 (D. Colo. Mar. 30, 2026)

‍

Why This Case Is Important

This landmark federal ruling establishes a crucial framework for how attorney work-product protections apply to artificial intelligence tools in litigation. It directly impacts how legal teams draft protective orders, manage data privacy, and select software when handling sensitive electronically stored information.

Overview Text

In Morgan v. V2X, Inc., the pro se plaintiff filed an employment discrimination lawsuit against corporate defendant V2X. Both sides used AI tools for litigation workflows. A discovery dispute arose when V2X moved to amend the existing protective order to restrict the plaintiff's AI use and compel disclosure of his specific AI software.

The plaintiff opposed, arguing V2X withheld an overdue insurance policy disclosure to force an amendment. He contended that restricting his analytical AI created an unfair technological gap, while his software choice was protected under the work-product doctrine.

The court resolved whether work-product protections extend to AI tool selection and how protective orders should govern AI usage. In doing so, the court distinguished this matter from precedents like United States v. Heppner, Warner v. Gilbarco, and Sporck v. Peil by addressing pro se advocacy, third-party cloud computing, and contractual data protection.

Ruling Summary

• Work-Product Protects AI Litigation Work
  The court held that Federal Rule of Civil Procedure 26(b)(3) protects AI-generated materials prepared for litigation. The judge emphasized that the importance of these protections is "magnified in the context of AI — one of the most powerful knowledge tools ever to become available to the masses." Rejecting arguments that using third-party AI waives confidentiality, the court asked: "Does that mean that anyone with a Gmail account has forfeited all rights to confidentiality and privacy?" The ruling takes care not to contradict U.S. v. Heppner, in which a defendant’s decision to use AI tools independent of their counsel is not protected work product.
  

• Disclosure of AI Platform Required
  Despite recognizing work-product protections, the court ordered the plaintiff to identify the specific AI software used to process confidential records. Tool selection alone does not inherently betray mental impressions without proof. As the ruling noted, "you have not demonstrated that identifying the tool itself will reveal your mental impressions or legal strategy." Therefore, V2X was entitled to learn the platform's name to assess potential data security risks.

• Strict Protective Order Standard Imposed
  The court drafted its own strict protective order language, barring the submission of confidential data to standard consumer AI models unless data processing agreements are in place. The court acknowledged that "this provision will (at least for now) bar the parties from using most, if not all, mainstream low-to-no-cost AI to process Confidential Information." This rigorous requirement ensures vendors are contractually prohibited from retaining or training models on protected discovery inputs. The ruling acknowledged that this restriction would be more harmful to pro se litigants.

Expert Analysis

Erin Corken, Esq., CEDS, FIP, AIGP, CIPP/US, CIPP/E, CIPM, Senior Solutions Engineer, Exterro

This case adds another voice to a growing chorus telling lawyers to use enterprise AI solutions with real contractual protections in place. That message echoes ABA Formal Opinion 512, which grounds this expectation in the ethical duty of confidentiality under Model Rule 1.6. Confidentiality requires keeping client information in enterprise tools with real, binding data protections, not in consumer platforms whose privacy settings the provider can change at will. Lawyers who skip this step aren't just risking an ethics complaint; they're building a case file on a foundation that opposing counsel, or a judge, may later be entitled to interrogate.  

‍

Tip Text

Audit your litigation technology stack immediately. Ensure any enterprise AI platforms have binding data processing agreements that prohibit vendor model training on uploaded records. Make sure your preservation playbook is up to date using the Exterro Modern eDiscovery Checklist.

‍