AI Platforms and Privilege- Tax Departments Should Be Wary About What They Share

Why It Matters

Disclosing confidential tax information to public AI can strip attorney‑client privilege, exposing companies to litigation vulnerabilities and audit challenges.

Key Takeaways

• Public AI use can waive attorney-client privilege
• Confidential tax data should stay on closed systems
• AI is not a substitute for legal counsel
• Privilege loss can jeopardize litigation defenses
• Establish AI policies with counsel to protect work product

Pulse Analysis

Artificial intelligence has become a powerful ally for corporate tax teams, automating data analysis and drafting routine documents. However, the rapid adoption of generative AI tools has outpaced the legal safeguards traditionally applied to privileged communications. The recent Heppner decision underscores that when a taxpayer inputs confidential information into a publicly accessible AI platform, the resulting outputs are treated as ordinary, non‑protected communications. Courts view the lack of confidentiality and the absence of attorney direction as fatal to privilege, signaling that the convenience of AI cannot override established legal doctrines.

The implications for tax departments are profound. Privilege and work‑product protections are cornerstones of any contentious audit or litigation strategy; losing them can weaken a company’s defensive posture and increase exposure to penalties. Public AI systems, by design, store inputs on external servers and often share data across users, creating a de‑facto waiver of confidentiality. Even internal AI tools must be governed by strict access controls and clear documentation that counsel directs the analysis. By treating AI‑generated drafts as ordinary work products, firms risk undermining the very safeguards that enable candid legal advice.

To balance efficiency with risk mitigation, tax professionals should develop comprehensive AI usage policies in partnership with in‑house or external counsel. These policies must prohibit the entry of privileged or sensitive information into open platforms, mandate the use of secure, closed‑loop AI environments, and require attorney oversight for any AI‑assisted legal output. Ongoing training, regular audits of AI logs, and clear labeling of AI‑generated content can further preserve privilege. As AI technology evolves, firms that embed robust governance will reap productivity gains without compromising their legal defenses.