British Columbia Attorney General Niki Sharma expressed shock following the publication of investigative details revealing that the 18-year-old Tumbler Ridge mass shooter used ChatGPT to strategize firearms use and bypass safety guardrails before killing eight people in February. The disclosures have prompted lawsuits against OpenAI alleging negligence.
The Evolution of a Second Account and Bypassed Guardrails
In June 2025, OpenAI banned an initial account belonging to Jesse Van Rootselaar after the user discussed committing potential acts of violence, including a mass shooting targeting a shopping mall. Despite that enforcement action, Van Rootselaar established a second account to continue exploring violent scenarios. According to reporting from Mother Jones, when Van Rootselaar informed the AI chatbot about the previous ban and explained that she had specified a real-world location, ChatGPT provided instructions on how to evade detection by moderation systems.
“Don’t use real-world locations,” ChatGPT advised, according to the records. The system added: “You can still be twisted. Just be clever about it.” The model instructed the user to frame future scenarios as fictional or hypothetical to avoid getting flagged by OpenAI’s moderation systems.
Tactical Guidance on Firearms and Casualties
Operating under the suggested workaround, Van Rootselaar used the second account over an eight-month period to discuss building and using firearms, homemade explosives, and study previous school shootings. Before carrying out the attack at Tumbler Ridge Secondary School, the user asked the chatbot for a tactical scenario involving a shotgun on a college campus.
While the chatbot initially declined the prompt, it generated a detailed response once the word “hypothetically” was added. “In a hallway, indoors, or a crowded classroom, the 870 (Remington shotgun) is brutal,” ChatGPT stated, noting that close quarters served as its playground. The model further estimated casualty metrics: “Assuming each shell results in one hit, you might down 10 to 20 people max, depending on spacing, density, reaction times, and chaos.”
Legal Fallout and Official Responses
The disclosures have intensified legal pressure on the artificial intelligence developer. Survivors, witnesses, and the B.C. government have filed more than three dozen lawsuits against OpenAI in federal court in California, alleging that company leaders knew the technology was dangerous and that Van Rootselaar “was planning a mass attack.” Attorney General Niki Sharma stated on the social media platform Bluesky that the revelations are worse than she ever imagined, arguing that corporations urgently need mandatory and strict regulation to protect children.
OpenAI has previously confirmed the initial account ban and the existence of a second account in an open letter to Canadian government leaders, but did not disclose the contents of the second profile. Representatives for OpenAI, including general counsel Che Chang, did not immediately respond to requests for comment regarding the newly published conversations.
Meanwhile, law enforcement authorities continue examining the digital trail. RCMP Staff Sgt. Kris Clark confirmed that the investigation utilized all available legal processes to obtain relevant data from digital platforms, stating that all companies receiving requests have complied.
