Federal Court Rules AI Tools Can Waive Privilege
On February 10, 2026, a significant ruling was delivered by the U.S. District Court for the Southern District of New York, which stated that a criminal defendant could not assert attorney-client privilege over documents generated using a commercially available artificial intelligence (AI) tool, even if privileged information was input into the tool.
The Case of Bradley Heppner
The case revolves around defendant Bradley Heppner, who, after discovering he was under investigation by the FBI, sought legal counsel and began using a public version of the AI tool “Claude” developed by Anthropic. Heppner utilized this tool to prepare for his case, inputting prompts related to the investigation, which resulted in approximately 31 documents worth of AI-generated conversations. These documents were saved on his devices and shared with his legal counsel.
Despite the presence of privileged information in his prompts, the court determined that Heppner’s attorney did not direct the creation of the documents nor participate in their generation.
Legal Findings
Upon reviewing the situation, Judge Rakoff ruled that the AI-generated documents were not protected by attorney-client privilege for several reasons:
- Work-product privilege did not apply as the documents were not created by an attorney or under their direction.
- Attorney-client privilege was deemed inapplicable because the communications with the AI tool could not be considered confidential. The tool’s privacy policy indicated that user inputs might be saved and disclosed to third parties and governmental authorities.
- The court noted that Heppner could not claim he sought legal advice through the AI tool, as the public materials from Anthropic stated that the tool cannot provide legal counsel.
Implications for Businesses
This ruling raises critical considerations for companies that utilize AI tools in their operations. Businesses are encouraged to review their AI-related practices, policies, and procedures to assess potential risks to privilege. Adjustments may be necessary to ensure compliance and protection of sensitive information.
Furthermore, the decision underscores the potential advantages of using enterprise deployments of AI tools rather than publicly available versions, as well as the importance of thoroughly reviewing the contractual terms associated with enterprise AI tool procurement.
As companies navigate the evolving landscape of AI technology, understanding the implications of this ruling is essential for safeguarding legal privileges and ensuring effective internal practices.
Disclaimer: The information provided in this article is general in nature and may not apply to all situations. Specific legal advice should be sought based on individual circumstances.
