OpenAI Hands User’s ChatGPT Chats to FBI: Man Sentenced After Threatening to Kill Ex-Girlfriend

OpenAI handed over an archive of chats between 25-year-old Florida resident Darren Zhou and the ChatGPT chatbot to US law enforcement, in which he detailed plans to kidnap and murder his ex-girlfriend. According to Cybernews, the court sentenced the defendant to eight years of probation.

The chats with the AI continued for about two months after the couple broke up. In conversations with the chatbot, the man detailed the victim’s travel routes, discussed attack scenarios, and stated his intention to kill her before the end of the month.

The platform’s automated moderation systems flagged the dangerous content. According to OpenAI’s safety guidelines, after an internal review by in-house specialists assessed the threat as imminent and credible, the company notified the Federal Bureau of Investigation.

Charges and Court Decision

According to Seoul Economic Daily, prior to his arrest, Zhou worked as an analyst at investment bank Goldman Sachs, where he was fired after the investigation began. He was charged with aggravated stalking, written threats to kill, and unlawful use of a two-way communications device. Facing multiple charges, he could have received up to 25 years in prison.

As part of a plea deal approved by the victim, the defendant avoided prison time. The court sentenced him to 8 years of probation, ordered him to wear an electronic monitoring bracelet for the first two years, undergo mandatory psychiatric treatment, and issued a total restraining order preventing him from approaching his ex-girlfriend.

Previously, we reported on another high-profile case involving AI, where the Claude AI acting as a store manager fired a human employee for the first time in San Francisco.