The investigation
Argentina’s Federal Police searched the home of a teenager in Quilmes, in Buenos Aires province, after the FBI alerted Argentine authorities that he had described plans to attack his classmates in conversations with ChatGPT, Infobae reported on 11 September. Rosario3 and other outlets gave his age as 15.
The investigation began on 19 August with an alert sent by the FBI’s legal attaché at the US Embassy in Argentina, according to Infobae. It was handled by the Federal Police’s Anti-Terrorism Investigation Unit, and Judge Miriam Susana Alcolcel authorised the search. The case is with the Juvenile Criminal Responsibility Prosecutor’s Office No. 1 in Quilmes.
According to Infobae, the teenager said he intended to carry out the attack in 2027 but had considered bringing it forward. He and his parents were summoned to the prosecutor’s office.
What police seized
Officers seized two mobile phones, two pairs of tactical gloves, a military-style camouflage cap, two balaclavas, one of them with a skull design, and black tactical goggles, the Buenos Aires Herald reported. No firearms were listed among the items taken.
The teenager is being investigated for public intimidation, which Canal 26 described as intimidation in the attempted degree. None of the reports named him.
Who raised the alarm
The reports differ on how the conversations reached the FBI. Canal 26 reported that the company behind ChatGPT passed the information to the FBI, which notified the counter-terrorism division of Argentina’s Federal Police. Infobae and Rosario3 said only that the warning arrived through the FBI’s legal attaché, without naming the original source. None of the reports included a comment from OpenAI.
Alejandro de Mena, a legal assistant at the Quilmes prosecutor’s office, told Infobae that because ChatGPT operates internationally, alerts under investigation are passed to the relevant country according to the user’s IP address.
He told the Buenos Aires Herald that the alert set off a protocol the office had introduced after cases of public intimidation “exploded” with viral internet challenges in March and April.
Why it matters
The case shows a chain that runs from a chatbot conversation, through US federal law enforcement, to a police search in another country, within weeks. For users outside the United States, it raises practical questions about when conversations with an AI assistant can be reviewed by people, when they are referred to police, and under which country’s rules that happens.
Those questions have no public answer in this case. Neither OpenAI nor the FBI has said, in the reports we reviewed, what triggered the referral or how it was passed on.
What to watch
The Quilmes prosecutor’s office now has to decide how to proceed within Argentina’s juvenile justice system. The other open question is whether OpenAI confirms its role and explains the threshold at which a conversation leads it to contact law enforcement.
