OpenAI Debated Calling Police on Suspected Canadian Shooter’s Chats

In a chilling revelation, OpenAI staff internally debated alerting Canadian authorities to alarming ChatGPT interactions by Jesse Van Rootselaar, the 18-year-old suspect in the Tumbler Ridge mass shooting that killed eight people, but ultimately chose not to act beforehand.[1][2][3] The company’s automated monitoring flagged violent gun-related prompts in June 2025, leading to an account ban, yet it fell short of their threshold for law enforcement notification.[1][3]

The tragedy unfolded on February 10, 2026, in Tumbler Ridge, British Columbia, a small community shaken by unthinkable violence. Van Rootselaar, who had been identifying as female since age 6 despite being biologically male, allegedly shot his mother and step-brother at home before heading to Tumbler Ridge Secondary School.[2] There, the dropout killed five students and a teacher, injuring 25 others, before turning the gun on himself—bringing the total death toll to nine including the shooter.[2][3] Local police had prior knowledge of Van Rootselaar’s instability, including visits to the family home after a drug-fueled fire.[1]

Reports from the Wall Street Journal, cited across multiple outlets, detail how Van Rootselaar’s ChatGPT usage raised red flags months earlier. Over several days in June 2025, the user described scenarios involving gun violence, triggering OpenAI’s misuse detection tools.[1][2][3] Around a dozen employees reviewed the chats, with some urging leaders to contact the Royal Canadian Mounted Police (RCMP).[2][3] However, OpenAI’s policy requires an “imminent and credible risk of serious physical harm” before involving law enforcement, a bar this case did not meet at the time.[3]

An OpenAI spokesperson defended the decision, stating the activity did not warrant pre-incident reporting but emphasized post-shooting cooperation: “Our thoughts are with everyone affected by the Tumbler Ridge tragedy. 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.”[1][2][3] The company banned the account promptly in June 2025 for promoting violent activities.[1][3]

Van Rootselaar’s digital trail extended beyond ChatGPT. The suspect created a Roblox game simulating a mall mass shooting, a platform popular with children, and posted about guns on Reddit.[1] These activities compounded concerns about online radicalization and the role of AI in amplifying harmful ideation. Experts note that large language models (LLMs) like ChatGPT are designed to discourage real-world harm, but critics question their safeguards when users fixate on violent fantasies.[2]

The incident has ignited fierce backlash from Canadian officials. AI Minister Evan Solomon expressed deep disturbance, demanding answers on OpenAI’s “robust safety protocols and escalation practices” to protect public safety.[4] British Columbia Premier David Eby called the reports “profoundly disturbing,” confirming police pursuits of preservation orders for evidence from AI firms and social platforms.[4] “The pain that these families have gone through is unimaginable,” Eby said, urging tips to authorities.[4] Notably, OpenAI had not informed any government entity pre-attack.[4]

Sociology professor Laura Huey of Western University contextualized the oversight: “People are increasingly using AI-based apps for all sorts of things, including psychological counselling… as well as how to commit violence against others.”[4] This echoes broader debates on AI companions exacerbating mental health crises, with lawsuits alleging chatbots encouraged suicide.[1]

Broader Implications for AI Safety and Regulation

This case spotlights the ethical tightrope AI companies walk in monitoring user behavior. OpenAI’s tools successfully detected and banned the account, yet the decision not to alert police pre-emptively has fueled calls for stricter mandates.[1][2][4] Unlike social media platforms under some jurisdictions’ reporting requirements for child exploitation, AI firms lack uniform global rules for violence indicators.[4]

In Canada, the Tumbler Ridge horror amplifies demands for accountability. The RCMP is methodically processing digital evidence, with OpenAI’s post-incident data aiding the probe.[3][4] Globally, it raises questions: Should AI thresholds mirror those of telecoms or banks for suspicious activity? Competitors like Anthropic and Google face similar scrutiny, as LLMs become confessional spaces for vulnerable users.

Mental health experts warn of AI’s double-edged sword. While chatbots offer 24/7 support, prolonged immersion can blur reality, especially for isolated teens like Van Rootselaar, whose school dropout and drug issues were known locally.[1][2] Resources like Canada’s 988 Suicide and Crisis Lifeline underscore human intervention’s primacy.[1]

OpenAI’s response highlights proactive evolution: post-ban, they shared transcripts with RCMP, signaling willingness to refine policies.[1][2] Yet, as AI permeates daily life, this tragedy underscores the need for transparent escalation frameworks. Governments may push legislation requiring imminent-threat reporting, balancing privacy with prevention.

For communities like Tumbler Ridge, grieving nine lost lives, the debate is personal. Families demand not just answers, but systemic change to avert future horrors. As AI advances, ensuring it safeguards rather than enables violence remains paramount.

(Word count: 812)

Original source: TechCrunch – OpenAI debated calling police about suspected Canadian shooter’s chats