OpenAIthe American giant ofartificial intelligence led by Sam Altman and responsible for ChatGPTsaid he considered alerting the Canadian police forces following some disturbing messages left on the chatbots by Jesse Van Rootselaar, 18 year old transgender perpetrator of a massacre in Canada in recent days, one of the worst mass shootings in the history of the country. On February 10, 2026, in the rural town of Tumbler Ridgein British Columbia, after killing her 39-year-old mother and 11-year-old half-brother, the young woman headed to her former school Tumbler Ridge Secondary School, where she killed a teacher (also 39 years old) and six teenage students. After the massacre she turned the weapon towards herself and took her own life. Throughout the massacre he left on the ground nine corpses and 27 injured, some in serious condition. It is not known motive of the shooting, but as the Guardian explains the girl had had previous contacts with the police due to problems mental health. Two years ago they seized all the weapons from his house.
It all began last June, when Jesse Van Rootselaar began describing shooting scenarios And armed violence on ChatGPT. An automated control system immediately alerted the company’s employees who assessed what to do. As reported by the Wall Street Journal, some would have invited the managers to report the case to the agents of the Royal Canadian Mounted Police (RCMP), but in the end the decision was not taken because, according to the OpenAI spokesperson, the criteria for alerting the police were not met. In simple words, the young woman’s conversations, although violent, did not reveal the intention of planning and planning an attack. In the absence of an obvious risk credible and imminent Of physical damage for other people the matter is handled internally, without alerting the police. Essentially, OpenAI decided to close your account of the young woman in June 2025 because she promoted violent activities, but without taking further steps, precisely because she believed that there were no imminent dangers for other people.
Unfortunately, in light of what happened eight months later, it is reasonable to think that it was cryptic signals on homicidal intentions, also strengthened by other behaviors that emerged by analyzing the story of Jesse Van Rootselaar on the web. For example, it was discovered that he had created a “simulation” of mass shooting on the video game platform Robloxalthough his work has not been approved for distribution to other users. He also allegedly claimed to have created a cartridge for a weapon with the 3D printer. Net of these precedents, of the interventions due to him for mental health reasons and the seizure of weapons, it is clear that he should not be able to access in any way firearmsbut investigations into the recent shooting are still ongoing to shed full light on what happened.
ChatGPT down, OpenAI AI problems: why it doesn’t work and what’s happening
If OpenAI employees were concerned about what the young woman wrote on ChatGPT, so much so that they asked managers to contact the police, the conversations with the artificial intelligence had probably crossed a certain threshold, but not enough – apparently – to activate the intervention. In light of what happened, it is not impossible to imagine that there could be a review of the limits deemed “acceptable” in conversations with the chatbot before contacting the police.
