By Mike Scarcella

June 23 (Reuters) - A federal judge has denied a bid by a business executive charged in a securities fraud prosecution to block a search warrant seeking data from ChatGPT maker OpenAI, rejecting his arguments that the request would sweep in protected communications with his defense lawyers.

U.S. District Judge Lorna Schofield in Manhattan on Monday lifted a temporary block on the search warrant in her ruling against Richard Kim, who was charged in August with defrauding investors in his crypto venture Zero Edge.

The warrant seeks records of his communications with a chatbot operated by OpenAI. Prosecutors say Kim’s accounts at the AI company and at a rival platform likely contain evidence of his alleged scheme to defraud investors of Zero Edge over how their funds would be used.

While the judge declined to quash the search warrant, Schofield’s ruling does not bar Kim from later arguing that prosecutors should be blocked from using any evidence they obtained from OpenAI.

Kim's lawyers did not immediately respond to a request for comment. He has pleaded not guilty to securities fraud and wire fraud charges.

OpenAI, which is not a party to the litigation, also did not immediately respond to a request for comment. The U.S. attorney’s office in Manhattan declined to comment.

The judge's order is part of a growing number of disputes over whether and when conversations with artificial intelligence systems can be shielded from prosecutors.

Kim’s dispute parallels a February decision by U.S. District Judge Jed Rakoff in Manhattan, who ruled in an unrelated criminal case that prosecutors could obtain a defendant’s communications with Anthropic’s chatbot Claude and that such exchanges were not protected by attorney-client privilege.

Kim’s lawyers had argued in a court filing that the warrant was overbroad and risked exposing attorney-client communications and other protected information after his arrest last year. They said he used AI tools to conduct “case-related research” that would reveal defense strategy if disclosed.

Prosecutors countered that any challenge to the warrant must come after the search.

The government also told Schofield that Kim’s communications with AI platforms are not privileged because chatbots are not lawyers, users cannot reasonably expect confidentiality with chatbots and such tools do not provide legal advice.

Prosecutors told the court that they are prepared “to respect individual assertions of privilege properly justified by defense counsel after their review of the data produced by the providers.”

The case is United States v. Richard Kim, U.S. District Court, Southern District of New York, No. 1:25-cr-00359-LGS.

For United States: Michael DiBattista and Ryan Nees of the U.S. Attorney’s Office for the Southern District of New York

For Kim: Kristoff Williams and Tamara Giwa of Federal Defenders of New York

Read more:

Judge rules both sides in lawsuit misused AI, disqualifies lawyers

Group fights Shell Oil bid for AI prompts used in climate lawsuit

OpenAI says ChatGPT is not a lawyer, asks court to toss insurer's lawsuit

AI errors in US murder case lead to discipline for Georgia prosecutor