Axios: AI Chatbot Logs Emerging as New Courtroom Evidence, Escalating User Privacy Risks

According to Observant Beating monitoring, in the Musk v. OpenAI case, OpenAI CEO Greg Brockman's personal diary was forcibly disclosed as evidence. Axios today reported that this type of privacy breach is rapidly spreading to the AI field: courts are increasingly admitting user conversations with chatbots as admissible evidence, and these conversations usually do not enjoy privileges similar to those of lawyers or psychologists.

This trend has been confirmed in several recent lawsuits. In February of this year, a U.S. federal judge ruled that a man's rehearsed conversation with Claude in preparation for a meeting with his lawyer could be used as evidence against him in a criminal case. Additionally, in murder cases in Florida and arson cases in Los Angeles, large-scale chatbot records have also appeared as courtroom exhibits.

Legal experts warn that AI chatbot records can be even more damaging to individuals than personal diaries. Diaries only provide a one-way record, while chatbots are designed to continuously ask questions and prolong conversations, making it easy to induce users to disclose details they would not otherwise have written. With complete interaction timestamps, chatbots can precisely capture a user's subjective cognition and intent at specific points in time. If exemptions for AI chats are not established in the future, this looming threat of potential disclosure at any time will cause a seismic shift in litigation.