A 15-Year-Old in Argentina Asked ChatGPT What He'd Need to Attack His Classmates. The FBI Flagged It, and Police Raided His Home.

BUENOS AIRES — Argentine Federal Police raided the home of a 15-year-old in Quilmes, a suburb south of the capital, after the FBI passed along an alert that the teenager had been using ChatGPT to plan an attack on his classmates. The raid, carried out Thursday on the order of Judge Miriam Susana Alcolcel, turned up two cell phones, two pairs of tactical gloves, a military camouflage cap, a skull-print balaclava, a second camouflage balaclava and black tactical goggles. No firearms were found.

The case began on August 19, when the FBI's legal attaché at the U.S. Embassy in Buenos Aires notified Argentine authorities that a user had been asking the chatbot what he would need to carry out an attack on his schoolmates. The conversations were surfaced by OpenAI's automated monitoring, which flags accounts that use certain terms associated with violence. The teenager's account was traced through his IP address, and the Federal Police's counter-terrorism investigations unit used open-source intelligence work to identify him and locate his home.

"ChatGPT operates internationally, and when investigating, alerts are generated based on the IP number of the device," said Alejandro de Mena, legal assistant at Quilmes Juvenile Criminal Prosecution Office No. 1, adding that "authorities then intervene in the relevant country's jurisdiction." He compared the mechanism to the way child-exploitation images are automatically reported across borders.

According to investigators, the teenager had told the chatbot he planned the attack for 2027, though he also discussed moving the date forward. Photos recovered during the raid showed him trying on the masks. He has been charged with "public intimidation" and summoned to the prosecutor's office. Neither he nor his parents have given statements. De Mena said public-intimidation cases "exploded" in Argentina in March and April, driven by viral internet challenges, and that this one arrived amid a series of similar raids and trials.

The episode is one of the first publicly documented cases of an AI company's safety monitoring leading directly to a police intervention in another country before any attack occurred. OpenAI has faced criticism for the opposite failure. In February, after an 18-year-old killed eight people in Tumbler Ridge, Canada, The Wall Street Journal reported that OpenAI had banned her account in June 2025 when monitoring flagged chats about gun violence, and that staff debated contacting Canadian police but decided the activity "did not meet the criteria for reporting to law enforcement." The company reached out to the Royal Canadian Mounted Police only after the shooting.

Argentina has its own recent history. In March, a 15-year-old opened fire at a school, killing one person and wounding eight, and in November 2025 a 16-year-old in the Caballito neighborhood of Buenos Aires was arrested with replica weapons and written attack plans after a separate FBI tip about his social media posts. The Quilmes case is the first in which the tip came from what a teenager typed into a chatbot.