Federal judge allows search of ChatGPT records in crypto fraud case
At a glance:
- A federal judge ruled prosecutors can access ChatGPT records of Richard Kim, former CEO of Zero Edge, in a $3.8M fraud case.
- The warrant covers Kim's OpenAI data from October 2023 to May 2026, including prompts and responses.
- The decision sets precedent on AI chat logs as legal evidence, rejecting claims of attorney-client privilege.
What happened
A federal judge in Manhattan has ruled that prosecutors can compel OpenAI to hand over ChatGPT account records belonging to Richard Kim, the former chief executive of cryptocurrency startup Zero Edge, as part of an ongoing fraud investigation. Kim is accused of diverting approximately $3.8 million of a $4.3 million fundraising round for personal crypto trades and online gambling. Following his arrest, court filings suggest Kim used ChatGPT to research his case, including trial strategy and prompts related to misappropriating funds.
US District Judge Lorna Schofield rejected the defense's attempt to block the warrant, determining that AI chat logs qualify as third-party digital evidence. The ruling permits prosecutors to obtain Kim's OpenAI records from October 2023 through May 2026, including account information, prompts, and responses. Kim has pleaded not guilty to securities and wire fraud charges. The case underscores how AI interactions are increasingly viewed as legally relevant data, akin to emails or search history.
Legal implications
The defense argued that Kim's ChatGPT conversations should be protected under attorney-client privilege, claiming the data contains internal thoughts and defense strategies. However, prosecutors countered that privilege requires confidential communication between a client and a licensed attorney for legal advice. Since AI tools like ChatGPT are not legal professionals, the court sided with the prosecution. Judge Schofield's ruling does not definitively resolve whether any of Kim's records are privileged but allows the warrant to proceed.
This case builds on a prior ruling in United States v. Heppner, where Judge Jed Rakoff found that exchanges with Anthropic's Claude chatbot were not protected by privilege or work-product doctrines. Rakoff emphasized that AI platforms act as third-party data collectors and noted the defendant had used Claude independently, not under attorney supervision. These decisions suggest courts may increasingly treat AI interactions as discoverable evidence, particularly when used without legal oversight.
Broader context
As AI tools become integral to research and decision-making, their role in legal proceedings is evolving. Users often treat chatbots as confidential advisors, but courts are clarifying that such interactions lack the protections of traditional legal communications. The Zero Edge case highlights risks for individuals using AI for sensitive tasks, including legal strategy, without considering data retention policies.
The ruling may influence future cases involving AI-assisted legal work. While AI use directed by attorneys could potentially receive different treatment, the current precedent emphasizes that third-party platforms are not bound by confidentiality agreements. Legal experts anticipate sharper distinctions will emerge as courts balance privacy expectations with law enforcement needs in the AI era.
Key takeaways
The case serves as a cautionary tale for AI users: conversations with tools like ChatGPT, Gemini, or Claude may not remain private in legal contexts. Companies and individuals should review AI platforms' data policies and consider how interactions might be interpreted in investigations. For OpenAI and similar firms, the ruling reinforces obligations to comply with lawful warrants while navigating user trust and privacy concerns.
The intersection of AI and legal discovery is likely to grow more complex. As the technology becomes ubiquitous, courts will need to address nuanced questions about intent, confidentiality, and the role of AI in professional and personal contexts. The Zero Edge case is a pivotal step in defining those boundaries.
FAQ
What did the judge rule regarding ChatGPT records in the crypto fraud case?
Why did the defense claim the ChatGPT records should be protected?
What precedent does this case set for AI and legal proceedings?
More in the feed
Prepared by the editorial stack from public data and external sources.
Original article