As artificial intelligence becomes more common in property management, concerns about privacy and legal discovery are growing with it. One recent federal court ruling has been interpreted by some commentators to mean that using an AI platform makes a company’s internal information publicly available or automatically discoverable.

That conclusion goes much further than the ruling itself. The case involved a specific dispute over attorney-client privilege and litigation work product, not a general declaration that every business conversation entered into an AI platform becomes public.

Still, the decision offers an important warning. Property managers should understand the difference between privacy, confidentiality, privilege, and discoverability before putting sensitive information into any AI tool.

What Happened in United States v. Heppner

In United States v. Heppner, a financial-services executive facing criminal charges had used Claude to explore issues related to his legal situation. The resulting materials were found on computers seized by the government, and his attorneys argued that the documents should be protected by attorney-client privilege or the work-product doctrine.

The court rejected those arguments. Claude was not the defendant’s attorney, and the documents had not been created as communications between the defendant and legal counsel. The court also focused on the circumstances surrounding their creation, including the involvement of a third-party AI platform and the platform’s applicable privacy terms.

The ruling did not establish that every AI conversation is automatically public. It found that the particular documents at issue were not protected from disclosure under the asserted legal doctrines. A summary of the ruling and its privilege analysis is available from Reuters.

That distinction matters because legal privileges are narrow protections. A document does not become privileged simply because it discusses a legal problem or is later forwarded to an attorney.

Privacy and Privilege Are Not the Same

Privacy refers generally to how information is collected, stored, used, and shared. A platform’s contract and privacy policy may promise that customer information will not be used for model training or disclosed except under defined circumstances.

Attorney-client privilege is a separate legal doctrine. It generally protects qualifying confidential communications between an attorney and client made for the purpose of requesting or providing legal advice.

The work-product doctrine is different again. It can protect certain materials prepared in anticipation of litigation by or for a party or its representatives, but its application depends on the circumstances.

A platform can offer strong commercial privacy protections without making everything entered into it legally privileged. Conversely, material may qualify for protection in litigation even though it was processed through a third-party system, depending on the facts and the safeguards in place.

Property managers should not use these terms interchangeably. A promise that information will not be used to train a model is valuable, but it is not the same thing as a court recognizing attorney-client privilege.

A Different Court Reached a More Nuanced Result

A later federal decision in Morgan v. V2X, Inc. demonstrates why broad conclusions about AI and discovery are premature. That case involved a self-represented plaintiff using an AI platform while preparing an employment lawsuit.

The court recognized that AI-assisted litigation preparation could potentially contain protected work product. It explained that routing information through a third-party system does not necessarily eliminate every expectation of privacy or automatically waive all applicable protections.

However, the plaintiff was required to identify the AI tool he used. The court also amended the case’s protective order to restrict the use of confidential discovery materials with AI platforms unless the provider offered specific contractual safeguards.

Those safeguards included restrictions on using inputs for model training, limits on disclosure to third parties, and the ability to remove or delete confidential information. The court’s detailed order illustrates that AI use is not categorically protected or categorically discoverable; the result depends on what was entered, why it was created, which protections applied, and how the platform handled the information. Read the court’s order in Morgan v. V2X, Inc..

Business Records Were Never Automatically Shielded

The discussion around these cases can create the impression that business records were confidential in litigation until AI arrived. That is not how discovery generally works.

Relevant emails, text messages, internal notes, contracts, accounting records, recorded calls, policies, and other business materials may already be subject to discovery when a company becomes involved in litigation. Whether a particular record must be produced depends on relevance, proportionality, the court’s rules and orders, applicable privileges, and the circumstances of the dispute.

Using AI does not necessarily create that underlying exposure. If an employee asks AI to summarize an email that is already relevant and discoverable, the original email did not become discoverable because AI processed it.

The AI interaction may create an additional record, however. That new record could contain prompts, summaries, conclusions, or other information that raises separate preservation and discovery questions.

This is why companies should reject both extremes. AI does not automatically broadcast every business record to the public, but AI conversations should not be treated as private notebooks that can never be accessed or produced.

The Greater Risk Is Poor Data Governance

For most property management companies, the immediate concern is not whether every AI prompt will appear in a lawsuit. The more pressing risk is employees entering sensitive information into tools without understanding how those tools store, process, or use it.

Property management companies routinely handle Social Security numbers, bank information, identification documents, screening reports, lease records, access instructions, medical accommodation information, and confidential owner data. That information should not be copied into a general-purpose AI platform merely because doing so is convenient.

The company should determine which AI products are approved, which account types employees may use, what data is prohibited, and whether information must be removed or anonymized before processing. It should also review retention terms, model-training settings, access controls, deletion options, security commitments, and procedures for responding to legal demands.

A free consumer account and a business or enterprise service may operate under different contractual terms. Property managers should confirm the terms that apply to their specific account instead of relying on assumptions about the platform as a whole.

Legal Matters Require Additional Caution

Employees should never treat an AI chatbot as a substitute for the company’s attorney. AI can produce inaccurate explanations, overlook jurisdiction-specific rules, or recommend a strategy without understanding the complete facts.

Using AI to research or organize information related to pending or anticipated litigation can also affect privilege, confidentiality, preservation duties, and protective orders. The safest approach is to consult counsel before placing case-related information into an AI system.

If an attorney approves the use of AI, the company should follow the attorney’s instructions regarding the platform, account type, information permitted, and handling of the resulting materials. It should not assume that forwarding an AI-generated document to counsel later will automatically protect the original interaction.

Court orders can also impose restrictions beyond a platform’s ordinary privacy terms. Confidential discovery received from another party may not be uploaded to an AI tool unless the protective order and provider contract specifically permit it.

Create an AI Policy Before an Incident Forces One

A practical AI policy should begin with approved uses and prohibited information. Employees need clear direction about whether they may use AI for resident correspondence, lease summaries, maintenance records, financial analysis, legal matters, screening information, or employee documentation.

The company should establish review requirements based on risk. A draft marketing email does not require the same controls as a fair housing response, legal notice, accommodation request, or communication about a resident’s account.

Access should also be managed at the company level whenever possible. Business accounts, centralized administration, single sign-on, appropriate retention settings, and employee training provide stronger control than allowing every team member to create an unrelated free account.

Finally, the policy must evolve. AI platforms change their features, contracts, privacy terms, and data controls regularly, so approval should not be treated as a permanent one-time decision.

The Lesson Is Caution, Not Panic

The emerging case law does not support the simple claim that using AI makes an entire business an open book. It does show that courts will apply established principles of confidentiality, privilege, work product, and discovery to new technology based on the specific facts before them.

Property managers do not need to abandon AI because of a sensational headline. They do need to understand what information they are entering, which platform is receiving it, what contractual protections apply, and whether the material relates to a legal dispute.

Responsible AI adoption requires more than choosing a tool. It requires governance, training, approved systems, careful handling of sensitive information, and coordination with legal counsel when privilege or litigation is involved.

The companies that build those disciplines can continue benefiting from AI without pretending that the technology is either completely risk-free or inherently unsafe.

This article provides general business information and is not legal advice. Privilege, confidentiality, and discovery rules vary by jurisdiction and depend heavily on the facts. Consult qualified legal counsel regarding your company’s AI policies or any pending legal matter.

Stay Informed with PMAssist Insider

AI is changing property management, but responsible operators need more than headlines to understand what those changes mean for their businesses.

Subscribe to PMAssist Insider for practical insights on artificial intelligence, technology, compliance, operations, and profitability.