ChatGPT conversations can be used as evidence in court

ChatGPT conversations are increasingly making their way into civil and criminal cases, where they can be obtained and used as evidence without the legal protections that cover privileged communications, The Washington Post reports.

The Post found at least 12 court cases involving chatbot conversations over the past two years. The real number is likely higher, since evidence obtained during litigation often remains outside the public record.

In one case, a teenager sued social media companies over alleged harm to his mental health. Defense attorneys obtained his ChatGPT history, including a conversation in which he discussed the possibility of receiving a million-dollar settlement.

Chatbot records are also appearing in criminal investigations. In Florida, OpenAI contacted the FBI after a user described plans to attack his former girlfriend. Authorities identified and arrested the man on stalking and electronic threat charges. He later pleaded guilty and received eight years of probation.

OpenAI uses automated systems to detect potentially dangerous behavior and may refer credible threats to law enforcement. The number of government requests for user data increased fourfold in the second half of 2025.

AI conversations are becoming relevant in civil litigation as well. A tire salesman in Michigan asked ChatGPT whether deleted Yahoo emails could be recovered through legal proceedings. The judge treated the exchange as evidence that he had attempted to conceal information and ordered him to pay a fine.

Legal experts interviewed by the Post expect cases involving chatbot histories to become more common as AI assistants are used for increasingly personal conversations. Unlike discussions with lawyers, doctors or other professionals covered by specific confidentiality protections, conversations with AI generally do not carry the same legal privilege, potentially making them accessible to courts and law enforcement when they become relevant to a case.