User Beware: AI Tools May Waive Attorney Client Privilege and are Not Protected by the Work Product Privilege

Parties to litigation, or potential litigation, should take great caution when using Artificial Intelligence (“AI”) tools as the information shared with an AI tool or provided by an AI tool may be subject to court ordered disclosure. Earlier this year, a federal judge granted a motion to overrule the alleged claims of privilege and work product on certain documents that were submitted by the defendant to an AI tool. As a result of this ruling, the party asserting claims of privilege to withhold information disclosed to an AI platform was ordered to produce these records.

The case at issue, United States v. Heppner, No. 25-cr-00503-JSR (S.D.N.Y. Oct. 28, 2025), involves a party who faces charges of securities fraud, wire fraud, and related offenses. At the outset of the case, federal agents seized numerous electronic devices and hard copy records from the defendant’s home.

In response to this seizure, the defendant claimed that the attorney client privilege and the work product doctrine prevented the government’s ability to review 31 documents contained on seized devices. This claim is founded upon the established legal principle that communication between a client and their attorney made for the purpose of obtaining or providing legal advice is protected from disclosure in discovery or other legal matters. This privilege ensures open communication, allowing clients to discuss matters honestly without fear of disclosure in legal proceedings. Also, that the work product of an attorney, including case strategies, be protected from discovery by an adverse party.

In Heppner, the defendant argued that the contested documents consisted of AI-generated material created for the “purpose of obtaining legal advice.” Among other things, the documents reflected that the defendant had shared communications from his counsel with an AI tool. They also reflected that the defendant sought to obtain advice and guidance from this AI tool and independent of his attorneys.

In rendering a decision, the Court noted that the AI tool used by the defendant “had an express provision that what was submitted was not confidential.” Moreover, the Court heard the defendant’s acknowledgement that the AI generated materials “were prepared by the defendant on his own volition”. This acknowledge stood in contrast to the “core purpose of Work Product Doctrine” framed by the Court, that it “is to protect the mental strategies of counsel in anticipation of litigation”. Ultimately, the attorney client privilege was held to be waived by the defendant and that the work product privilege did not apply.

As a result of this ruling, the government was able to view all disputed records, including communications with counsel, prompts to the AI tool regarding strategy, reports generated outlining legal strategies, and other information that traditionally would not be shared with an adversary in litigation.

Though this decision comes in the context of criminal litigation, the holding is equally relevant to non-criminal matters and should come as a stark reminder to parties that are involved in litigation or may be involved in litigation. AI tools are not to be utilized without acknowledgement of the significant risks that are faced. In civil litigation, a party may seek discovery regarding the use of AI tools, and you may be compelled to produce records of AI prompts, reports, and other materials. Prompts or other records provided to an AI tool may be construed as an admission of a party, especially if there is a prompt requesting a legal response to some alleged or actual wrongdoing. More importantly, AI chatbots, secretaries, note taking, and other systems used while on a call with an attorney could waive any claim of privilege, and risk undermining your claims or defenses.

While users can protect themselves by using closed system, enterprise level tools with strong privacy protections, there is no guarantee that the information disclosed to or received from an AI tool will be afforded the protections that currently exist for clients seeking legal guidance or the development of legal strategies with an attorney. Individuals and businesses alike should be mindful of the risk of AI and adopt policies and procedures tailored to that risk. Those policies and procedures should be prepared with counsel familiar with issues of privilege and the work product doctrine. If those policies and procedures are not in place and if litigation is a potential, then counsel should be involved immediately to ensure that the appropriate privileges apply to any investigation, response, or strategy.