Plan told to ChatGPT reached the police: Farmer arrested in Brazil

In Brazil, a 36-year-old farmer was arrested by the police after disclosing details of a severe crime plan against his own child through his correspondence with ChatGPT. The incident also drew attention to how safety systems operate on artificial intelligence platforms.

OpenAI forwarded the message to US authorities

The incident took place near the city of São Gabriel da Palha in Espírito Santo state. According to the police, the suspect had been corresponding with ChatGPT for nearly two months, leaving information about plans for violence against his child.

OpenAI detected these communications from a safety perspective and forwarded the information to the US Federal Bureau of Investigation (FBI). Subsequently, the FBI alerted local law enforcement agencies through the CyberLab division of the Brazilian Ministry of Justice.

The suspect was captured one day before the plan was to be carried out

Brazilian police stated that the suspect was apprehended on June 19, one day before the date the plan could have been executed. Information regarding the real threat to the child's life was provided to law enforcement authorities in advance.

The police announced that he is suspected under charges related to threats, incitement to crime, and attempted murder. The identity of the suspect has not been disclosed.

Importantly, in this case, ChatGPT did not serve as a tool for committing a crime, but rather as a digital environment where the safety system worked to help detect dangerous communication.

This event has further intensified debates on how interactions with artificial intelligence are monitored from a security standpoint and the procedures by which information is handed over to law enforcement agencies when a serious threat is detected.