News

US Man Arrested After OpenAI Reports Him to FBI for Sharing Ex-Girlfriend Murder Plot with ChatGPT

US Man Arrested After OpenAI Reports Him to FBI for Sharing Ex-Girlfriend Murder Plot with ChatGPT
안내

We only offer this video
to viewers located within Korea
(해당 영상은 해외에서 재생이 불가합니다)

▲ ChatGPT

An American man who confessed his plot to murder his ex-girlfriend to the artificial intelligence chatbot ChatGPT has been caught following a report by OpenAI.

Darren Zhou, a 25-year-old financial analyst at a major investment bank living in Florida, was arrested after detailing his plan to kill his former partner during conversations with ChatGPT in March, according to IT media outlet Futurism and local newspaper The Palm Beach Post on the 14th (local time).

He sent messages to ChatGPT such as, "I will kill her by the end of this month," and "If I can't have her, no one can."

He also recorded the victim's hobbies and frequently visited places, and expressed jealousy toward other men.

Law enforcement authorities explained that his messages went beyond a simple expression of emotion, reaching the level of a detailed crime plan—mentioning plans to wait with flowers in the parking lot of the fitness facility his ex-girlfriend attends, shoot and kill her if rejected, and then take his own life.

He also bragged to ChatGPT that he had purchased rifles, handguns, and shotguns, and sent photos of them to his ex-girlfriend.

OpenAI detected Zhou's harmful conversation pattern and reported the chat history to the Federal Bureau of Investigation (FBI).

Right after his arrest, he was released on a 100,000-dollar bond, but was fired from his investment bank job.

During the subsequent trial, he pleaded guilty, but through a plea agreement, his sentencing was deferred, and he received eight years of probation along with two years of wearing an electronic monitoring bracelet.

Zhou's defense attorney claimed, "He is not a violent person nor a bad person, but merely suffered from mental health difficulties," adding that his condition has now improved.

The defense also explained that, contrary to the messages sent to ChatGPT, he did not actually own firearms.

OpenAI previously faced controversy for failing to report to authorities after learning that a perpetrator of a mass shooting in British Columbia (BC), Canada, had repeatedly mentioned specific crime scenarios to ChatGPT months before the attack.

Consequently, families of the mass shooting victims have filed a damage lawsuit against OpenAI, and the government of British Columbia has also announced legal action, including retaining legal representation.

(Photo: AP, Yonhap News)
※ Please note: This article was translated by AI and may contain errors.
Copyright Ⓒ SBS. All rights reserved. 무단 전재, 재배포 및 AI학습 이용 금지

Most Read