In the aftermath of one of Canada's deadliest mass shootings, new details have emerged about the suspect's interactions with artificial intelligence tools in the months leading up to the tragedy. Jesse Van Rootselaar, the 28-year-old identified as the gunman in the February 10 attack at Tumbler Ridge Secondary School in British Columbia, had engaged in conversations with OpenAI's ChatGPT that described violent scenarios involving firearms. These exchanges, which occurred last June, triggered internal alarms at the company, prompting some employees to urge leaders to notify law enforcement. However, OpenAI ultimately decided against doing so, citing a lack of imminent threat.
The shooting at the rural high school in the small mining community of Tumbler Ridge left nine people dead and 27 others injured, including Rootselaar, who died of an apparent self-inflicted gunshot wound at the scene. It marked the deadliest mass shooting in Canada since the 2020 Nova Scotia attacks that claimed 22 lives. Witnesses described chaos as gunfire erupted during the school day, with students and teachers barricading themselves in classrooms while emergency services rushed to the scene. Local authorities, including the Royal Canadian Mounted Police (RCMP), have been investigating the motive, which remains unclear, though Rootselaar reportedly had no prior criminal record in the area.
According to a report by The Verge, Rootselaar's ChatGPT interactions began raising red flags within OpenAI's monitoring systems. The conversations involved detailed descriptions of gun violence, which the AI's automated review flagged as potentially harmful. Several OpenAI employees, concerned that these posts could signal real-world intent, pushed for the company to alert authorities. Despite these internal discussions, OpenAI's leadership opted not to proceed, determining that the content did not meet the threshold for an "imminent and credible risk" of harm.
OpenAI spokesperson Kayla Wood addressed the decision in a statement to The Verge, explaining the company's rationale. "While the company considered referring the account to law enforcement, it was ultimately decided that it did not constitute an 'imminent and credible risk' of harm to others," Wood said. She added that a review of the interaction logs showed no evidence of active or imminent planning for violence. As a precautionary measure, OpenAI banned Rootselaar's account, but no further actions, such as notifying police, were taken at the time.
The revelations have sparked questions about the responsibilities of AI companies in monitoring user behavior. OpenAI, the maker of ChatGPT, has faced increasing scrutiny over how it handles potentially dangerous content generated or discussed on its platforms. In this case, the company's internal debate highlighted the tension between user privacy and public safety. Employees who spoke anonymously to The Verge expressed frustration, believing that earlier intervention might have prevented the tragedy, though Wood emphasized that hindsight should not overshadow the complexities of such decisions.
Wood also noted OpenAI's commitment to balancing these priorities. "OpenAI’s goal is to balance privacy with safety and avoid introducing unintended harm through overly broad use of law enforcement referrals," she said. The spokesperson reiterated that the company does not routinely report user interactions unless they clearly indicate immediate danger, a policy designed to protect users from unwarranted surveillance.
Following the shooting, OpenAI took steps to cooperate with investigators. "Our thoughts are with everyone affected by the Tumbler Ridge tragedy," Wood stated. "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." This post-incident outreach contrasts with the pre-shooting inaction, underscoring the challenges in predicting violent acts from online activity.
Tumbler Ridge, a town of about 2,500 residents nestled in the foothills of the Rocky Mountains, has been grappling with grief since the attack. The school, which serves students from grades 8 to 12, was a central hub for the community. Interim Superintendent Maria Gonzalez described the impact in a press conference the day after the shooting: "Our hearts are broken. This was a place of learning and growth, now forever changed by unimaginable loss." Counseling services have been made available, and a memorial has sprung up outside the school gates, adorned with flowers, candles, and messages of condolence.
Investigators have pieced together a timeline of the day. The attack began around 10:30 a.m., with Rootselaar entering the school armed with a semi-automatic rifle legally purchased in Alberta. He had moved to Tumbler Ridge just a year earlier, working odd jobs in the local coal mines. Neighbors described him as reclusive, with few close ties, though some recalled heated arguments in the months prior. The RCMP has not confirmed a specific motive but said Rootselaar left behind a manifesto-like document ranting against societal pressures, though its contents remain sealed.
The incident has reignited national debates on gun control in Canada, where strict laws already limit firearm access compared to the United States. Prime Minister Justin Trudeau visited the community last week, vowing to review mental health resources and online radicalization. "We must do more to protect our children," Trudeau said during a vigil. Advocacy groups like the Canadian Coalition for Gun Control have called for tighter regulations on semi-automatic weapons, pointing to this as the latest in a string of school shootings.
From a technological standpoint, the case illustrates the limitations of AI safety measures. ChatGPT, powered by OpenAI's GPT models, includes safeguards to detect and respond to violent prompts, often refusing to engage or warning users. However, Rootselaar's interactions apparently skirted these boundaries, allowing detailed discussions to proceed until the automated flag. Experts in AI ethics, such as those from the Alan Turing Institute, have noted that while companies like OpenAI invest heavily in moderation—spending millions on human reviewers—the sheer volume of interactions makes proactive threat detection difficult.
One anonymous OpenAI employee told The Verge that the internal push to report Rootselaar stemmed from patterns in the conversations that echoed known indicators of violence, such as fixation on weapons and isolation. Yet, without concrete plans, the decision to hold back was upheld. This mirrors broader industry practices; competitors like Google and Meta have similar policies, reporting only when legally required or when threats are explicit.
As the RCMP's investigation continues, questions linger about what more could have been done. Community leaders in Tumbler Ridge, including Mayor Sarah Jenkins, have expressed mixed feelings about the AI angle. "It's heartbreaking to think words on a screen might have foreshadowed this," Jenkins said in an interview. "But we can't blame technology alone—our society needs to address the root causes of such despair."
Looking ahead, OpenAI has indicated it may refine its reporting protocols in light of the tragedy. The company updated its statement on February 21, adding the commitment to ongoing support for law enforcement. For the victims' families, however, such measures come too late. A GoFundMe campaign has raised over $500,000 for funeral costs and rebuilding efforts, a small solace in the wake of profound loss.
The Tumbler Ridge shooting serves as a stark reminder of the intersection between digital interactions and real-world consequences. As AI tools become ubiquitous, the balance between innovation and oversight will remain a critical challenge. For now, the community mourns, piecing together lives shattered in an instant, while policymakers and tech leaders grapple with lessons from a preventable horror.