• Use of enterprise versions of AI tools, while no guarantee of maintaining privilege, should bolster privilege claims.
• To bolster work product claims, clients and other non-lawyers using AI tools at the direction of counsel to assist with a legal case should make it clear in their prompts that they are acting at the direction of counsel.
• Privilege logs should clearly denote the basis for the privilege and that the AI tool was used with the expectation of confidentiality. The entry should also state that the communication reflects legal advice from an attorney or that the work was done at the direction of counsel, or both.

On February 10, 2026, Judge Rakoff of the U.S. District Court for the Southern District of New York ruled from the bench that documents a client created using a commercial generative AI tool and sent to his lawyer were not protected by privilege. Defendant Bradley Heppner was arrested on charges of securities and wire fraud on November 4, 2025. During the search of his mansion, federal agents seized electronic devices containing approximately thirty-one documents generated using Anthropic’s AI tool Claude. After he received a grand jury subpoena and had engaged legal counsel, Heppner used Claude to prepare reports outlining his defense strategy and potential legal arguments.

While Judge Rakoff has not yet issued a written opinion, the decision has significant implications for protecting client communications that involve the use of AI tools.

BACKGROUND

After the arrest and seizure of Heppner’s electronic devices, Heppner’s counsel claimed privilege over the Claude documents, describing them on a privilege log as “[a]rtificial intelligence-generated analysis conveying facts to counsel for [the] purpose of obtaining legal advice.” Defense counsel asserted that the AI reports were a means of consolidating Heppner’s thoughts for the express purpose of communicating with counsel. However, Heppner’s counsel conceded that Heppner prepared the AI documents on his own initiative, not at his counsel’s direction.

The government moved for a ruling that the documents were not protected by privilege, arguing that the AI-generated documents were not confidential and not created for the purpose of obtaining legal advice. The government further contended that the defendant could not retroactively cloak unprivileged documents by later transmitting them to counsel. In addition, the government argued that the work product doctrine did not apply because the materials were prepared by the defendant on his own initiative and the doctrine “does not protect a layperson’s independent internet research.”

During a pre-trial conference, Heppner’s counsel from Quinn Emanuel argued that the documents were protected by the attorney-client privilege because they incorporated information conveyed to their client during the course of the representation. With respect to the work product doctrine, Heppner’s counsel argued that a report created by a client in anticipation of litigation should be protected even if it was not done at the direction of legal counsel.

THE COURT'S DECISION

Judge Rakoff issued an oral ruling that neither the attorney-client privilege nor the work product doctrine protected the AI-generated documents. The decision rests on traditional principles of privilege.

The attorney-client privilege protects (1) communications, (2) among only privileged parties, (3) made for the purpose of providing or obtaining legal advice. Importantly, the protection of the attorney-client privilege is lost if the communication is shared outside of the privileged parties. The party claiming privilege has the burden of showing that confidentiality was maintained. Judge Rakoff stated that the attorney-client privilege did not apply because the communications were shared with a third-party tool that did not maintain confidentiality.

Second, Judge Rakoff held that the work product doctrine did not protect the documents. The work product doctrine protects (1) legal work product, (2) discussing legal strategy, (3) prepared by or at the direction of legal counsel, (4) in anticipation of litigation. Judge Rakoff rejected Heppner’s arguments that the work product doctrine could apply because the AI-generated reports did not reflect the legal strategy of Heppner’s legal counsel, although they contained theories generated by the client and Claude. Since neither Heppner nor the AI tool are legal counsel, and Heppner was not working at the direction of Heppner’s legal counsel, the materials were not protected by the work product doctrine. Judge Rakoff noted that the AI tool’s disclaimer that users have no expectation of confidentiality also undermined the work product doctrine claim.

KEY TAKEAWAYS FOR PROTECTING PRIVILEGE AND BROADER IMPLICATIONS FOR LEGAL RESEARCH

The decision is the first we are aware of that finds that use of a consumer (non-enterprise) AI tool with otherwise potentially privileged materials may result in a loss of privilege. Both the Department of Justice and Judge Rakoff take the view that sharing information with consumer AI tools is inconsistent with the requirement to keep privileged communications confidential. In its motion, the DOJ noted that the applicable Anthropic Privacy Policy provides that prompts can be used to train their model and inputs may be disclosed to governmental regulatory authorities and third parties. Judge Rakoff ruled from the bench that the defendant “disclosed it to a third-party, in effect, AI, which had an express provision that what was submitted was not confidential.” Although not entirely clear from the hearing transcript, we believe that Judge Rakoff should view the use of an enterprise AI tool (which does not train on inputs and maintains confidentiality of inputs) differently, such that the use of an enterprise tool with otherwise privileged materials should not result in loss of privilege. For that reason, use of enterprise versions of AI tools, while no guarantee of maintaining privilege, should bolster privilege claims.

To bolster work product claims, clients and other non-lawyers using AI tools at the direction of counsel to assist with a legal case should make it clear in their prompts that they are acting at the direction of counsel. Courts have found that the work product doctrine can protect AI-generated content where the prompts and use of the tools fulfill the criteria to assert privilege. The government’s brief suggested that there could be instances where the work product doctrine protects content a client generates using AI, such as if “counsel directed the defendant to run the AI searches.”

This publication is for general information purposes only. It is not intended to provide, nor is it to be used as, a substitute for legal advice. In some jurisdictions it may be considered attorney advertising.