Judge rules that LLM provided legal advice is open to discovery [pdf] (storage.courtlistener.com)

A recent ruling by a U.S. District Court in New York has determined that documents created by a defendant using an artificial intelligence tool do not qualify for attorney-client privilege or the work product doctrine. The case revolves around Bradley Heppner, who allegedly created around thirty-one documents through the AI platform Claude, developed by Anthropic, concerning a criminal investigation against him. The court found that since the AI tool is not an attorney and Heppner did not obtain legal advice in the traditional sense, the documents are not privileged. This decision holds significant implications for the legal and AI/ML communities, as it clarifies the limitations of using AI in legal contexts. The ruling underscores that communications generated through AI cannot retroactively be deemed privileged merely by sharing them with legal counsel. This case sets a precedent regarding the use of AI tools in legal scenarios and may influence how such technologies are utilized in legal research and case preparations, highlighting the need for clear understanding and guidelines about AI-generated content in legal proceedings.