Particle.news

Man Arrested in Espírito Santo After ChatGPT Conversations Revealed Plot to Kill Son

The case shows AI platform safety alerts can trigger international police action that leads to preventive arrests.

Overview

  • Police in São Gabriel da Palha carried out a preventive arrest on June 19 after investigators used files from AI conversations to obtain search and seizure warrants.
  • OpenAI flagged the user’s ChatGPT exchanges to the FBI, which forwarded the material to Brazil’s Ministry of Justice and then to Espírito Santo investigators.
  • Recorded messages told the AI the suspect planned to kill his 8-year-old son to avoid paying child support, said investigators who also report the suspect mentioned a firearm, rope and cyanide.
  • Authorities say the conversations include an attempted R$50,000 hire of a hitman and threats of wider attacks on schools and public sites; seized phones and other items are now undergoing forensic analysis.
  • The detainee admits making the searches but denies intent to carry out the killing and investigators are considering charges such as threat, attempted homicide and incitement pending perícia and further inquiry.