Resources

On February 10, 2026, Judge Jed Rakoff of the U.S. District Court for the Southern District of New York issued what appears to be one of the first rulings of its kind addressing whether materials created using a consumer AI platform can be protected by attorney-client privilege or the work product doctrine. In an oral ruling from the bench, Judge Rakoff held that documents generated by a criminal defendant using a commercial AI tool and later shared with counsel were not privileged, signaling that courts may treat consumer AI platforms as third parties for the purposes of determining the confidentiality of attorney-client communications.

The ruling sends a clear message that entering sensitive legal information into publicly available AI tools may jeopardize privilege protections, suggesting that even when AI-generated information is ultimately sent to an attorney a court may find that confidentiality of the information was waived because it was first shared with a third-party AI platform. As a result, even material that was explicitly created with the intent to assist counsel could still become discoverable in litigation, regulatory proceedings, or investigations if an AI platform was used to generate the material.

Judge Rakoff’s decision also emphasizes that the work product doctrine, which protects legal preparation and strategy developed in anticipation of litigation, may not apply when AI-generated materials are created independently by a client rather than at the direction of counsel. In this context, if an individual uses AI tools on their own initiative to analyze potential defenses or legal theories, that information may be treated as personal research rather than protected legal preparation under the work product doctrine.

The significance of this decision also lies in its treatment of consumer AI tools as potential third-party disclosures. The court relied in part on the terms and disclaimers of the AI platform that was used, which stated that user inputs were not confidential, reinforcing the view that submitting privileged information into an AI platform may undermine claims of attorney-client privilege or work product protection. The court did not directly address the use of commercial “enterprise” AI platforms, which typically operate under agreements restricting access to user data to ensure confidentiality. However, the reasoning suggests that future privilege determinations may turn on whether an AI platform functioned as (or was intended to function as) a confidential extension of counsel’s work product or as an outside third party that received voluntary disclosures from a client or other related party.

This first-of-its-kind ruling represents an early judicial signal that the convenience of generative AI carries meaningful legal risk. Until clearer judicial standards develop, the safest approach is to treat consumer AI tools as external third parties for confidentiality purposes and to always involve legal counsel before using them in connection with any legal matter, dispute, or investigation.

If you have questions about how this ruling may impact your organization’s use of AI tools, please contact BMD Attorney Jeff Joseph at jajoseph@bmdllc.com.