Court Shields ChatGPT Conversations From Discovery Demand in High-Stakes Lending Dispute

A New York judge has drawn a significant line around the privacy of artificial intelligence conversations, blocking an attempt to force OpenAI to hand over a litigant’s ChatGPT records in a commercial lawsuit.

The ruling emerged from a legal battle involving private lender Alpha Tech Lending and its founder, who sued former company president John Recchio III and others over allegations including breach of contract and unfair competition connected to a rival business venture. Recchio has denied the claims.

At the center of the dispute was a subpoena seeking extensive information from OpenAI about Recchio’s use of ChatGPT. The plaintiffs wanted access to prompts, uploaded documents, generated responses, and other materials they believed could shed light on the factual basis of Recchio’s legal arguments and defenses.

Justice Rhonda Fischer of the Nassau County Supreme Court rejected that request, concluding that the AI-generated exchanges were protected under legal work-product principles. The decision, which became public last week, is among the earliest judicial examinations of whether conversations with AI platforms can be compelled in civil litigation.

Recchio, who is representing himself, argued that the subpoena reached far beyond legitimate discovery. He contended that the request sought access to his private litigation workspace, including legal research, strategy discussions, draft arguments, and materials prepared in anticipation of court proceedings.

The court agreed, finding that the protections traditionally afforded to litigation-related work extend to AI-assisted research and preparation. In reaching that conclusion, Fischer referenced a federal ruling from Colorado that compared a self-represented litigant’s interactions with AI tools to confidential strategic work product. That court observed that while AI providers may retain user data, such retention does not automatically erase expectations of privacy or waive legal protections.

The plaintiffs’ legal team has indicated it intends to challenge the ruling on appeal.

Recchio welcomed the decision, arguing that civil discovery should not become an indirect route for opponents to obtain sensitive AI records, business information, or account data without a compelling legal basis and close judicial oversight.

OpenAI, which is not a party to the underlying lawsuit, did not publicly comment on the decision.

The ruling arrives as courts across the United States grapple with the growing role of generative AI in legal practice. Judges have increasingly been asked to weigh questions involving AI-generated research, attorney reliance on chatbots, and the extent to which user interactions with these systems should remain private.

As artificial intelligence becomes a routine tool in litigation preparation, the Nassau County decision may serve as an early marker in a broader debate over where discovery rights end and digital legal strategy begins.

Print Friendly, PDF & Email
Scroll to Top