Artificial intelligence now plays a role in how many family businesses analyze information, prepare for and record board and family meetings, and communicate with advisors. In 2026, courts are deciding what happens to that work in litigation, including whether an opposing party can obtain AI prompts and outputs, and whether using a public AI tool gives up privilege or trade secret protection. Shareholders have also begun testing whether AI-related decisions can support fiduciary duty claims against directors. The early answers are mixed. Owners, officers and directors who understand the risks now can help their companies capture AI's benefits while protecting confidential information, preserving privilege and making sound, defensible decisions.
The Legal Landscape Is Developing
Courts have reached different results on whether interactions with AI tools are protected from disclosure, so businesses should not assume that anything submitted to an AI platform is shielded. Recent decisions illustrate the tension: United States v. Heppner (S.D.N.Y. February 2026) found that a criminal defendant’s exchanges with a publicly available AI platform, made on his own rather than at his lawyers’ direction, were protected by neither the attorney-client privilege nor the work product doctrine, while Warner v. Gilbarco (E.D. Mich. February 2026), Morgan v. V2X (D. Colo. March 2026) and Assini v. Hayward (N.Y. Sup. Ct. June 2026) recognized work product protection for AI materials that self-represented litigants prepared for their own cases. Tate Group Automotive, LLC v. Legacy Automotive Capital, LLC (Tex. Bus. Ct. June 2026) applied the work product doctrine to a nonlawyer’s chat history with a publicly available AI platform, but only allowed some portions of the chats to be withheld on those grounds. None of these decisions involved a company using AI under a negotiated enterprise agreement, and the Heppner court suggested the analysis could differ where counsel directs the AI use. The law remains unsettled and may develop differently across courts and different jurisdictions. For now, the safest course is to treat AI interactions as potentially discoverable and handle them accordingly.
Key Risks for Business Owners, Officers and Directors
Fiduciary Duty and AI-Assisted Decision-Making
Officers and directors owe a duty of care that requires informed decision-making. When AI assists with financial analysis, strategic planning, hiring, risk assessment or other business decisions, the responsible officer or director remains accountable for the outcome. Relying on AI-generated content without meaningful human review could support a claim that the decision-maker failed to exercise the care expected of the position. Shareholder derivative complaints filed in 2026 against officers and directors of Adobe, Microsoft and Nvidia allege that they breached their fiduciary duties by approving, or failing to oversee, AI strategies that relied on allegedly unlicensed copyrighted training data, and by making public statements about those strategies that the plaintiffs contend were misleading. These cases are still being developed and litigated, with no court yet ruling on the merits of those allegations. Those suits target AI developers rather than companies that use AI in their operations, but each points to a lack of documented board-level oversight of AI, a gap that is common in some businesses with informal board practices. Commentary suggests that AI oversight when deployed is part of the board’s broader duty of care. The business judgment rule (a legal doctrine that shields good-faith, informed business decisions from being second-guessed by courts) will continue to protect directors who exercise genuine, informed judgment. But that protection covers only judgment that is exercised, and a documented record is the best proof that it was.
Discovery Risks: AI Prompts and Outputs
AI-related documents, such as prompts and outputs, are subject to the same discovery rules as other business records. In litigation, the opposing party may request these documents. In Assini v. Hayward, plaintiffs subpoenaed OpenAI directly for a self-represented defendant’s prompts, inputs and outputs, although the court ultimately quashed the subpoena on work product grounds (a doctrine protecting materials prepared in anticipation of litigation). In Tate v. Legacy, the court ordered production of certain pages of a nonlawyer’s AI chats that, after an in camera review, were found not to be work product, although the court otherwise protected the remaining chats and required that party to identify all discovery materials it had shared with ChatGPT. In Morgan v. V2X, the court rejected the litigant’s argument that the identity of his AI tool was protected work product and required him to disclose any AI platform he had used with information the defendant had designated confidential. Companies should extend document-retention and litigation-hold policies to expressly cover AI-related documents. Anything entered into or generated by an AI tool may need to be preserved and produced. Failing to do so can create avoidable spoliation (the loss or destruction of evidence) and credibility risks.
Confidentiality, Privilege and Trade Secret Risks
Entering sensitive company information into a third-party AI platform can create a serious risk of losing privilege or other protections. In Heppner, the court concluded that a defendant’s exchanges with a publicly available AI platform were not protected by the attorney-client privilege because the platform was not an attorney, the communications were not confidential and the provider’s privacy policy permitted it to use the data for training and to disclose it to third parties. Other courts have taken a more protective view of AI materials prepared by self-represented litigants and nonlawyer parties, but those decisions do not eliminate the need for caution. A federal court has also dismissed a trade secret claim based on disclosure to an AI provider. In Trinidad v. OpenAI (N.D. Cal. January 2026), the pro se plaintiff alleged that she developed her claimed secrets using ChatGPT, which required sharing them with OpenAI, a party under no duty of confidentiality, and the court dismissed her Defend Trade Secrets Act claim, holding that she had not alleged reasonable measures to keep them secret, a requirement for any trade secret claim.
Consumer-grade tools generally do not guarantee confidentiality. Before entering privileged, proprietary or trade secret information, confirm the provider’s contractual limits on storage, training, human review and third-party disclosure. The Morgan court built similar limits into its protective order: no party may enter confidential information into an AI platform unless the provider is contractually prohibited from storing it, using it to train its model or disclosing it to third parties (except as needed to deliver the service), and the party can have it deleted on request.
AI Is a Tool Worth Using, but Use It Responsibly
AI can deliver real business value: greater efficiency, stronger analytical capability and a competitive advantage for companies that use it thoughtfully. The goal is not to avoid AI. It is to use AI with the same care applied to any other business tool that touches sensitive information or influences important decisions. As the court observed in Morgan v. V2X, AI is “one of the most powerful knowledge tools ever to become available to the masses.” The question for business leaders is not whether to use AI, but how to use it in a way that protects the company and supports sound judgment.
Recommendations
The following practical steps can help address the key risk areas discussed above.
Compliance Keys
- Adopt a written AI governance policy and framework covering which tools may be used, for what purposes and who is responsible for reviewing AI-generated output before it is relied upon. Ensure through that policy and framework that AI-assisted work product is reviewed and verified before it is used by an officer or director in a business decision.
- Treat AI governance as part of the board’s oversight responsibilities. Directors should understand what AI tools the company is using, how those tools interact with sensitive information and what safeguards are in place. Record that oversight in board minutes.
Preserve Privilege, Confidentiality and Trade Secrets
- Do not input confidential information into an AI platform unless the platform is contractually prohibited from storing or using the input to train its model, or disclosing it to others, and the user can delete its information upon request.
- Retain written documentation of the affirmative steps the company took to secure confidentiality before entering information into the platform.
- Consult legal counsel before implementing AI policies or deploying AI tools that will interact with confidential, privileged, trade secret or regulated information.
- Be mindful that AI chat history may be discoverable in litigation and apply the same best practices to AI chats as you would to email and other document creation.
Conclusion
The law around AI in business is developing rapidly, and the standards that will ultimately govern this area are still being written. Business owners, officers and directors who adopt responsible AI practices now will be better positioned to capture the technology’s benefits and defend their decisions if questions arise later.
Contact Kristina Munsters, Scott Carvo or a member of our Private Client and Family Office Industry Group to discuss how AI governance fits into your company’s broader risk-management strategy.
Kristina Munsters is a partner in Warner’s Litigation and Dispute Resolution Practice Group, where she co-chairs Warner’s Business Breakup and Ownership Disputes subgroup and regularly defends officers and directors against breach-of-fiduciary-duty claims. She also advises clients on document retention, information governance and eDiscovery strategy.
Scott Carvo is a partner in Warner's Litigation and Dispute Resolution Practice Group, where he counsels clients on the responsible deployment of generative AI in business operations and litigation workflows and leads large-scale eDiscovery engagements, including in bet-the-company matters. He also represents companies in restrictive-covenant and trade-secret disputes and frequently speaks on AI governance and AI-assisted litigation.
