Federal Judge Rules AI Tool Queries Not Protected by Attorney-Client Privilege
Key Takeaways
- ▸A federal judge ruled that documents created using Anthropic's Claude AI tool are not protected by attorney-client privilege or work product doctrine, marking the first such judicial determination
- ▸The court found that using a public AI platform with a privacy policy allowing data collection and third-party disclosure eliminated reasonable expectations of confidentiality
- ▸The ruling was highly fact-specific, focusing on Claude's terms of service that permit Anthropic to collect user inputs/outputs, use data for training, and share information with government authorities
Summary
In a precedent-setting decision, U.S. District Judge Jed S. Rakoff of the Southern District of New York ruled that documents generated using third-party AI tools are not protected by attorney-client privilege or work product doctrine. The February 10, 2026 ruling in U.S. v. Heppner involved 31 documents that defendant Bradley Heppner created using Anthropic's Claude AI assistant, which contained information from his legal counsel.
Judge Rakoff rejected privilege claims on multiple grounds. He determined that Heppner had disclosed confidential information to a public AI platform with diminished expectations of privacy. The court noted that Claude's privacy policy explicitly states that Anthropic collects both user inputs and AI outputs, uses this data for training, and reserves the right to share information with third parties including government authorities. The judge also found the documents did not qualify as work product since Heppner created them independently, without direction from his attorneys.
The ruling appears to be the first judicial determination that interactions with publicly accessible AI tools containing privileged information are not themselves privileged. Legal experts suggest this is a highly fact-specific decision that may have limited broader applicability, particularly given the focus on Claude's specific privacy policy terms. However, the decision raises important questions about how traditional privilege protections apply in the age of AI assistance.
The case underscores growing concerns about confidentiality when using consumer AI tools for sensitive legal or business matters. Companies and legal professionals are being advised to carefully review the terms of service and privacy policies of any AI platforms they use, particularly regarding data retention, training practices, and third-party disclosure rights.
- Legal experts warn that companies and attorneys must carefully review AI tool privacy policies before using them for privileged or confidential communications
