Artificial intelligence has quickly become a useful tool for businesses, individuals, attorneys, accountants, and tax professionals. AI can help organize information, summarize documents, identify issues, and assist with research and drafting.
But when confidential tax, financial, or legal information is entered into an AI platform, an important question arises: Does using AI put confidentiality or legal protections at risk?
Two federal court decisions from February 2026 illustrate why the answer is not always straightforward. In United States v. Heppner, a federal court in New York concluded that certain communications with a publicly available AI chatbot were not protected by attorney-client privilege or the work-product doctrine. In Warner v. Gilbarco, Inc., a federal court in Michigan reached a different conclusion regarding AI-assisted litigation materials and work-product protection.
Although neither case involved international tax planning specifically, the decisions provide important lessons for individuals and businesses whose legal and financial affairs involve sensitive information.
Why AI Creates New Confidentiality Questions
Attorney-client privilege generally protects confidential communications between an attorney and client made for the purpose of obtaining or providing legal advice.
The work-product doctrine provides a separate protection for certain materials prepared in anticipation of litigation.
Neither protection automatically applies simply because a document contains legal or tax information.
Confidentiality matters.
When sensitive information is entered into a third-party AI platform, the platform may process, store, or otherwise handle that information. Depending on the service and its terms, information entered into an AI system may potentially be accessible beyond the individual using the tool.
That creates a fundamental legal question: Has the person using the AI system shared confidential information with a third party in a way that undermines the protection they were relying on?
The answer may depend on the specific AI platform, how it was used, who directed its use, what information was entered, and the legal protection being claimed.
What Happened in United States v. Heppner?
In United States v. Heppner, a defendant used a publicly available version of an AI chatbot while preparing materials related to a criminal investigation and potential legal defense.
The defendant argued that the resulting materials should be protected by attorney-client privilege or the work-product doctrine.
Judge Jed Rakoff of the U.S. District Court for the Southern District of New York rejected those arguments.
Among other things, the court emphasized that the AI platform was not an attorney and therefore could not establish an attorney-client relationship. The court also found that the communications were not confidential because they had been shared with the AI provider.
The decision is significant because it demonstrates that simply using an AI tool to discuss a legal matter does not automatically create or preserve legal protection.
In fact, the use of a publicly available AI system can raise the same type of confidentiality concerns that arise when confidential information is voluntarily disclosed to another third party.
Another Federal Court Reached a Different Result
Just days before the written Heppner decision, a federal magistrate judge in Michigan addressed AI-generated materials in Warner v. Gilbarco, Inc.
The case involved a pro se plaintiff who used generative AI tools in connection with employment litigation. The defendants sought discovery concerning the plaintiff's AI use.
The court declined to require production of the materials, finding that the work-product doctrine could protect the plaintiff's litigation-related materials and mental impressions. The court emphasized that generative AI programs are tools and that work-product waiver generally requires disclosure to an adversary, or disclosure in a manner likely to place the information in an adversary's hands.
The contrasting decisions demonstrate an important point:
Using AI does not automatically destroy every legal protection—but neither does the use of AI automatically preserve one.
The circumstances surrounding the AI use matter.
Attorney-Client Privilege and Work Product Are Not the Same
One of the most important lessons from these cases is that attorney-client privilege and work-product protection involve different legal principles.
Attorney-client privilege is generally focused on confidential communications made for the purpose of obtaining or providing legal advice.
Work-product protection is generally concerned with materials prepared in anticipation of litigation and the protection of an attorney's or party's litigation strategy and mental impressions.
As a result, an AI disclosure could raise different questions under each doctrine.
The Heppner and Warner decisions demonstrate why businesses and professionals should not assume that one broad rule applies to every AI interaction.
What Does This Mean for Tax and International Tax Matters?
For individuals and businesses with complex domestic or international tax matters, the stakes can be particularly high.
Tax planning frequently involves highly sensitive information, including:
- Personal and business financial records
- Tax returns and supporting documents
- Foreign account information
- Information concerning foreign entities and trusts
- Ownership structures
- Investment information
- Estate and succession plans
- Business transactions
- Cross-border transactions
- Information about family members and beneficiaries
- Communications concerning potential tax disputes or examinations
Uploading that information to a consumer AI platform without first understanding how the platform handles the information could create unnecessary confidentiality and security concerns.
For someone with international assets or business interests, the issue can become even more complicated because the information may involve multiple jurisdictions and multiple layers of legal and regulatory obligations.
AI Should Not Be Treated Like a Confidential Legal Advisor
One of the clearest lessons from Heppner is that an AI chatbot is not a lawyer.
An individual may ask an AI system to explain a tax provision, analyze a transaction, or suggest a legal strategy. But the fact that the conversation involves sophisticated legal or tax questions does not create an attorney-client relationship.
AI can be a powerful technology tool, but it does not replace the professional relationship between a client and qualified legal counsel.
This distinction is particularly important when dealing with complicated tax planning, cross-border transactions, estate structures, or international compliance matters.
What Should Businesses and Individuals Consider Before Using AI?
AI does not necessarily need to be eliminated from professional workflows. Instead, businesses should establish thoughtful controls around how the technology is used.
Understand the AI Platform
Before entering confidential information into an AI system, understand how the platform handles user information.
Review its privacy policies, data-use practices, retention policies, security measures, and applicable terms of service.
Not every AI platform provides the same level of confidentiality or security.
Avoid Entering Highly Sensitive Information Into Public AI Tools
A publicly available consumer chatbot should not automatically be treated as a confidential environment.
Businesses and individuals should be particularly cautious about entering:
- Tax return information
- Bank account information
- Passport or identification information
- Social Security numbers
- Confidential business records
- Privileged attorney communications
- Information about pending legal disputes
- Proprietary business information
- Details about trusts, estates, or family wealth structures
Establish Internal AI Policies
Businesses that use AI should consider creating clear policies addressing what employees may and may not enter into AI platforms.
An effective policy can address approved AI tools, prohibited information, security requirements, review procedures, and employee responsibilities.
Consult Your Attorney Before Using AI for Sensitive Legal Matters
If a matter involves potential litigation, a tax controversy, an IRS examination, a significant transaction, or another sensitive legal issue, it may be appropriate to discuss AI use with legal counsel before entering information into an AI system.
The circumstances surrounding how an AI tool is used can matter to the confidentiality analysis.
The Issue Is Especially Important for Cross-Border Clients
For internationally mobile individuals, multinational businesses, and families with assets in multiple countries, confidentiality is only one part of the equation.
Cross-border tax planning may involve information subject to different privacy, reporting, and regulatory requirements in different jurisdictions.
For example, a client's U.S. tax planning may involve information about foreign corporations, partnerships, trusts, bank accounts, investments, or family members living abroad.
That information can be extremely sensitive.
Before incorporating AI into a cross-border tax or legal workflow, businesses and individuals should consider not only whether the technology is useful, but also where their information is going, who may have access to it, how long it may be retained, and what protections apply.
What These Court Decisions Mean Going Forward
The law surrounding AI, confidentiality, privilege, and discovery is still developing.
The different outcomes in Heppner and Warner demonstrate that courts may analyze AI use differently depending on the circumstances. Other courts are continuing to address questions concerning AI-generated materials, prompts, confidentiality, and work-product protection.
For businesses and individuals, the safest approach is not to assume that an AI platform is automatically confidential simply because it is being used for a legal, tax, or business purpose.
Instead, AI should be incorporated into professional workflows deliberately and with appropriate safeguards.
Protecting Confidential Information in an Increasingly AI-Driven World
Artificial intelligence can provide meaningful benefits for businesses and individuals, but those benefits need to be balanced against confidentiality, privacy, and legal risks.
The recent federal decisions involving United States v. Heppner and Warner v. Gilbarco serve as an important reminder that traditional legal principles continue to apply even as technology changes.
For clients dealing with international tax planning, cross-border transactions, estate planning, asset protection, or complex business structures, protecting sensitive information should remain a central consideration.
At Bridgepointe Global Counsel, we help individuals, families, and businesses navigate complex U.S. and international tax, legal, and wealth-planning issues. Our approach integrates tax planning, corporate structures, estate planning, asset protection, and cross-border considerations to develop strategies tailored to each client's circumstances.
If you have questions about protecting confidential information while using AI or need guidance regarding a complex domestic or international tax matter, contact Bridgepointe Global Counsel to discuss your situation with experienced legal counsel.
This article is provided for general informational purposes only and does not constitute legal or tax advice. The application of attorney-client privilege, work-product protection, privacy laws, and other legal protections depends on the specific facts and circumstances.