OpenAI identified an account of the teenager who committed a violent attack, but did not alert the police, considering that it was not an imminent danger.

The American company OpenAI confirmed that the 18-year-old teenager, who killed eight people in British Columbia, Canada, had previously discussed with ChatGPT and showed signs of violent behavior. In June 2025, OpenAI detected Jesse Van Rootselaar's account through its automated monitoring systems for discussions about violence, but deemed it did not pose an imminent risk to alert authorities.

The tragedy occurred last week when the young girl first killed family members and then attacked a school, resulting in the deaths of five students. Although the girl had a history of mental health issues and had begun gender transition, the motives for the attack remain unclear. OpenAI subsequently collaborated with the police to provide information about the user's activity on the platform. The incident raises questions about the responsibility of AI platforms in preventing violence and collaborating with authorities to protect the public.