March 2, 2026


"I Told You So": Privacy Pitfalls in Using GenAI for Legal Matters Exposed by Recent Court Ruling

In a recent decision that echoes concerns raised in a previous analysis, the Southern District of New York delivered a stark reminder about the confidentiality—or lack thereof—regarding communications with GenAI platforms. The case, *United States of America v. Bradley Heppner*, highlights critical misunderstandings about privacy in the digital age, particularly in legal contexts.

Judge Rakoff's ruling clarified that discussions between Bradley Heppner and the GenAI platform, Claude, were not protected under attorney-client privilege or as work product. Heppner had used Claude to strategize for his defense, unknowingly compromising his privacy and legal strategy by not involving his lawyer directly in these communications.

The court pointed out that for communication to be covered under attorney-client privilege, it must be exclusively between the client and their lawyer, intended to be confidential, and for the purpose of obtaining legal advice. Since Claude is neither a lawyer nor a confidential medium—its terms of use clearly state that data can be used for training purposes—Heppner's interactions with the platform were deemed unprotected.

Moreover, the ruling emphasized that work product privilege, typically reserved to protect a lawyer's mental impressions and legal strategies, does not extend to AI-generated content created without direct supervision or instruction from an attorney.

The implications of this ruling are significant, not just for clients but also for lawyers who might consider using public GenAI tools for case strategy. Even if a lawyer directly inputs information into an AI like Claude, the confidentiality of this data is questionable if the platform retains the right to use it for other purposes, potentially waiving any privilege.

This case serves as a cautionary tale about the risks associated with using GenAI tools in legal matters without proper safeguards. Clients and lawyers alike should be wary of the terms of service for any platform they use and consider the potential discoverability of any information shared with AI.

As the legal profession continues to integrate more technology into its practices, both lawyers and their clients must stay informed about these developments and their implications for privacy and confidentiality. Understanding and navigating the terms of use of technological tools is not just advisable; it is essential for maintaining the integrity and confidentiality of legal work.