ChatGPT helped prevent a woman's possible murder
In America, OpenAI's security system helped law enforcement agencies prevent the murder of a woman.
ChatGPT helped prevent a woman's possible murder.
According to media reports, Darren Shida Zhou, a 25-year-old Goldman Sachs employee, consulted ChatGPT to plan the kidnapping, rape, and murder of his ex-girlfriend. He also planned to commit suicide after committing the crime.
OpenAI's security system automatically identified the correspondence as potentially dangerous. After that, OpenAI experts reviewed it, and it was they who decided to pass on the information to law enforcement agencies. ChatGPT did not independently report the crime to the police.
After receiving the information, the FBI and local law enforcement agencies began an investigation. As a result, Zhou was arrested.
The investigation also revealed that the man had repeatedly stalked his ex-girlfriend after the breakup, using different phone numbers and accounts.
Zhou pleaded guilty to three criminal charges: threatening to kill or cause serious bodily harm; stalking with a serious threat; and unlawful use of a communication device to commit a crime.
The judge sentenced him to eight years of probation without issuing a formal criminal sentence.
The case is one example of how artificial intelligence systems can be used to detect the threat of violence early. At the same time, it has once again raised the issue of the limits of privacy in users’ correspondence with artificial intelligence.