Photo: PAIGE TAYLOR WHITE / AFP / Getty Images
OpenAI knew something was wrong months before a mass shooter killed eight people at a British Columbia school — but the company never called police.
According to a report by The Wall Street Journal, around a dozen OpenAI employees were aware of deeply troubling ChatGPT interactions made by Jesse Van Rootselaar — the 18-year-old who carried out the February 10 massacre in Tumbler Ridge, British Columbia — as far back as June 2025. The interactions, which were first flagged by an automated review system, involved violent scenarios and discussions of gun violence that unfolded over several days. Despite those red flags, OpenAI chose not to alert law enforcement.
Van Rootselaar shot and killed his mother and half-brother at their home before heading to Tumbler Ridge Secondary School, where he fatally shot five students and an education assistant. Twenty-five others were injured. The shooter later died of an apparent self-inflicted gunshot wound.
OpenAI said in a statement that the shooter's account was banned in June 2025 for violating its usage policies, but that the company determined the posts did not meet its threshold for a law enforcement referral. The company's policy requires posts to show "an imminent and credible risk of serious physical harm to others" before authorities are contacted. OpenAI said it did not identify "credible or imminent planning" at the time.
Some employees reportedly pushed to go to police, but the company held firm. OpenAI cited privacy concerns and warned that over-reporting to law enforcement can cause "distress" to a young person and their family, and that doing so too frequently "can introduce unintended harm."
OpenAI did reach out to the Royal Canadian Mounted Police (RCMP) — but only after the shooting took place. In a statement, the company said: "We proactively reached out to the Royal Canadian Mounted Police with information on the individual and their use of ChatGPT, and we'll continue to support their investigation."
RCMP Staff Sgt. Kris Clark confirmed Friday (February 20) that OpenAI had contacted investigators following the shooting. He said "digital and physical evidence is being collected, prioritized, and methodically processed" as part of a thorough review of the shooter's online activity and electronic devices.
RCMP Deputy Commissioner Dwayne McDonald said investigators are still working to determine the origins of one of the two firearms used in the attack — a weapon that had never been seized by police and remains unaccounted for. Former RCMP member Chris McBryan, who once served with the Canadian Firearms Program, said determining where that weapon came from — and whether others were involved in obtaining it — would likely be a central focus of the continuing investigation.
Police had visited Van Rootselaar's home on multiple occasions in the past due to various incidents, and were aware of his mental health struggles. OpenAI said it regularly reviews its policies on alerting law enforcement, though the company did not say whether this incident would prompt such a review.