Upload of Documents to Anthropic Products Leads to Denial of Motion to Seal Record

Much has been written about whether use of A.I. waives the attorney-client privilege or work product protection. Tara Emory and Maura Grossman Address Protection of Attorney Prompts in “GenAI TAR” (Jul. 5, 2026); Deponent’s Use of A.I. to Answer Deposition Questions Barred; ChatGPT Was Not an Attorney (May 2, 2026); Important A.I. Work Product and Protective Order Decision (Mar. 31, 2026);  “Two Courts, Two Answers: When Does Using AI Waive Privilege?” (Mar. 3, 2026).

A court recently held that an upload to Claude waived any right to seal documents filed in court.

In Fry v. Fry, 2026 WL 2531956 (E.D. Penn. Aug. 27, 2026), the petitioner alleged that his wife had wrongfully removed their children from Germany to the U.S.  Petitioner also sought sealing of certain documents that had been filed.  They contained “personal identifiers such as social security numbers, dates of birth; financial account numbers; names of minor children; names and contact information of third-party military and medical personnel mentioned in his email correspondence and military records; prescription and medical records of Petitioner; correspondence with German officials; and details of domestic disputes between Petitioner and Respondent.”

After analyzing both the First Amendment and common law, and after in camera review, the court granted, in part, and denied, in part, the motion to seal.  It wrote in pertinent part:

With Petitioner’s renewed motion, the Court required Petitioner to submit an AI certification describing any assistance of AI that Petitioner utilized to prepare his motion…. Petitioner disclosed to the Court that he had uploaded his redacted documents to non-enterprise versions of Claude and Claude CoWork, products by third-party company Anthropic, in preparing his motion. As such, the Court finds that the full sealing of any document is not appropriate. Documents and information uploaded to Claude are processed on Anthropic’s servers, and Anthropic’s data privacy and data training pages make clear that such information is not private. See, e.g., Privacy Policy, Anthropic https://www.anthropic.com/legal/privacy (last accessed August 24, 2026); How Do You Use Personal Data in Model Training?, Anthropic, https://privacy.claude.com/en/articles/10023555-how-do-you-use-personal-data-in-model-training (last accessed August 27, 2026); see also United States v. Heppner, 820 F. Supp. 3d 292, 296 (S.D.N.Y. 2026) (holding that communications with Claude are not privileged due, in part, to Claude’s privacy policy). As such, it would be inappropriate for the Court to seal a document that has already been provided to a private, for-profit research and artificial intelligence company. [emphasis added].