OpenAI vows safety policy changes after Tumbler Ridge shooting

Following criticism from Canadian officials, OpenAI has pledged to strengthen its safety measures after failing to report a ChatGPT account linked to the suspect in the February Tumbler Ridge shooting to police. The suspect, Jesse Van Rootselaar, had an initial account banned in June 2025, seven months before the attack that killed eight people. OpenAI acknowledged that Van Rootselaar created a second account that bypassed their detection systems. In a letter to Canadian officials, OpenAI stated it has since changed its reporting protocols, making them more flexible and enlisting mental health experts to assess cases. The company claims that under the new guidelines, the suspect's activity would have been reported to law enforcement.

Canadian officials are investigating OpenAI's knowledge of Jesse Van Rootselaar's chatbot use prior to a mass shooting in Tumbler Ridge, British Columbia, this month. Van Rootselaar killed eight people before taking her own life. OpenAI had suspended Van Rootselaar's ChatGPT account in June, eight months before the attack, due to policy violations flagged by their abuse detection system. Canada's AI minister, Evan Solomon, has summoned OpenAI representatives to a meeting in Ottawa to discuss the company's safety protocols and decision not to alert authorities about the concerning messages, citing a lack of credible or imminent planning. The investigation aims to understand what OpenAI knew and why they didn't share the information with law enforcement before the tragedy.

Canada's artificial intelligence minister has summoned OpenAI representatives to explain why the company didn't alert police after suspending the account of Jesse Van Rootselaar in June 2023, prior to the February 2024 school shooting in Tumbler Ridge where she killed eight people. Van Rootselaar had described violent scenarios involving guns to ChatGPT, prompting the suspension for "furtherance of violent activities." OpenAI, however, determined the activity didn't represent "credible or imminent planning" and didn't notify Canadian law enforcement. The minister seeks clarification on OpenAI's safety protocols and escalation thresholds for involving police. The meeting comes as Canada considers regulating AI chatbots, particularly regarding minors' access. OpenAI stated they contacted the RCMP after learning of the shooting.

Canada's artificial intelligence minister has summoned OpenAI representatives to explain why the company didn't alert police after suspending the account of Jesse Van Rootselaar in June 2023, prior to the February 2024 school shooting in Tumbler Ridge where she killed eight people. Van Rootselaar had described violent scenarios involving guns to ChatGPT, prompting the suspension for "furtherance of violent activities." OpenAI, however, determined the activity didn't represent "credible or imminent planning" and didn't notify Canadian law enforcement. The minister seeks clarification on OpenAI's safety protocols and escalation thresholds for involving police. The meeting comes as Canada considers regulating AI chatbots, particularly regarding minors' access. OpenAI stated they contacted the RCMP after learning of the shooting.

In June 2025, OpenAI banned a ChatGPT account belonging to Jesse Van Rootselaar, the suspect in the February 12, 2026, mass shooting in Tumbler Ridge, British Columbia, where eight people were killed and 27 injured. OpenAI's abuse detection identified the account as potentially furthering violence, but the company did not alert authorities, deeming the usage below the threshold of a credible or imminent threat. Following the shooting, OpenAI contacted Canadian police with information about Van Rootselaar, who died by suicide at the school. The Wall Street Journal reported internal debate at OpenAI regarding alerting authorities prior to the event. OpenAI maintains its policy of only alerting authorities in cases of imminent risk to avoid unintended harm, and is reviewing the incident for improvements to its referral criteria.

OpenAI identified the account of Jesse Van Rootselaar in June 2023 for violating its usage policy due to concerns about violent activities. Van Rootselaar was later identified as the shooter in the Tumbler Ridge Secondary School shooting in British Columbia, Canada, which resulted in eight deaths. OpenAI considered alerting the Royal Canadian Mounted Police (RCMP) about the account but decided against it, as the activity did not meet the company's threshold for imminent and credible risk of serious physical harm. The account was banned in June 2023. Following the shooting, OpenAI contacted the RCMP with information about Van Rootselaar's use of ChatGPT and offered support for the investigation.