Judge Permits Search of ChatGPT Logs in $3.8M Crypto Fraud Case

For legal professionals, this case redefines the boundaries of digital discovery. The court’s decision that ChatGPT logs are akin to search history or emails signals that attorney-client privilege will not extend to AI-driven communication, no matter how confidential the user believes it to be.

The ruling by U.S. District Judge Lorna Schofield marks a pivotal moment in the intersection of AI, privacy, and criminal law. By allowing prosecutors to serve a search warrant on OpenAI for the ChatGPT records of Richard Kim, a former crypto startup CEO, the court has established that AI chatbot conversations are not inherently private and can be treated like any other third-party digital evidence. The case centers on allegations that Kim diverted $3.8 million of a $4.3 million Zero Edge fundraising round into cryptocurrency trading and online gambling. Following his arrest, Kim used ChatGPT to research his legal situation and trial strategy, generating prompts that, according to prosecutors, even touched on how to misappropriate investor funds. This unusual fact pattern underscores a growing challenge: as individuals increasingly turn to AI for advice—legal, financial, or otherwise—they may be unwittingly creating detailed digital records that can be used against them.

“

The case centers on allegations that Kim diverted $3.8 million of a $4.3 million Zero Edge fundraising round into cryptocurrency trading and online gambling.

The warrant, covering Kim’s OpenAI account from October 2023 through May 2026, seeks all prompts, responses, and account information. The defense argued that these records should be shielded under attorney-client privilege, as they revealed Kim’s thoughts and trial preparation. However, Judge Schofield rejected that claim, emphasizing that attorney-client privilege requires a confidential communication between the client and a licensed legal professional for the purpose of seeking legal advice. An AI chatbot, no matter how sophisticated, is not a lawyer. This ruling makes clear that users cannot assume that conversations with AI are protected; they are more akin to search queries or social media posts, which are routinely subpoenaed.

The decision is part of a broader trend of courts grappling with digital evidence. Emails, texts, and cloud-stored data have long been subject to search warrants, and now AI interactions join that list. This has significant implications for the AI industry. OpenAI and other providers may face increased pressure to clarify their data retention policies and to educate users that their chats can be accessed by law enforcement. If users fear that their conversations with ChatGPT might one day be read in a courtroom, they may avoid asking certain questions, limiting the tool’s utility.

For the cryptocurrency and startup communities, the case is a cautionary tale about the digital trail left by AI. Crypto traders and startup founders often rely on AI for market analysis and strategy; this precedent means those conversations are discoverable in investigations. The allegations of fund misappropriation also highlight the risks in startup funding, where investor money can be diverted if proper controls are lacking. The fact that Kim allegedly used ChatGPT to research his own fraud case—and perhaps even to brainstorm the fraud itself—adds a surreal layer to the investigation. Prosecutors now have a window into the defendant’s mindset, potentially documenting intent and premeditation.