A February 2026 federal court ruling in United States v. Heppner found that a fraud defendant's conversations with Anthropic's Claude AI were not protected by attorney-client privilege or work-product doctrine. Judge Jed Rakoff ruled that Claude is not a lawyer, has no confidentiality obligation, and Anthropic's terms of service explicitly allow data collection and disclosure to third parties — meaning the defendant had no reasonable expectation of privacy. The ruling contrasts with two other cases where pro se litigants' ChatGPT conversations were deemed work product, though legal analysts note key factual differences. In response, over a dozen major US law firms have issued advisories warning clients to treat public AI platforms as non-confidential environments and to use only private, closed AI deployments for anything legally sensitive.
Sort: