British Columbia has sued OpenAI and CEO Sam Altman over the Tumbler Ridge school shooting, turning a devastating loss of life into a major test of how AI companies respond to violent threats. The province alleges that OpenAI’s failure to warn law enforcement about flagged ChatGPT conversations helped leave a deadly danger undisclosed.
The lawsuit, filed on September 22, 2026, claims that the school killings could have been prevented if OpenAI had alerted police about Jesse Van Rootselaar’s use of ChatGPT. It accuses the company and Altman of negligence, while seeking damages and an order that would change how conversations linked to possible violence are handled.
What the lawsuit says OpenAI knew
Van Rootselaar’s interactions with ChatGPT were first flagged by OpenAI’s safety team in June 2025 after conversations about gun violence. OpenAI recommended contacting police, but Altman and other company leaders overruled that advice, according to the lawsuit.
Police did not receive an alert after the conversations were flagged. OpenAI deactivated Van Rootselaar’s account, but she was able to create a new account and continue using the platform. The province argues that these decisions created a failure point at the moment when human authorities may have been able to intervene.
OpenAI has stated that its systems flagged the account because they identify misuse in furtherance of violent activities. The company also stated that the issues did not meet its internal criteria for reporting to law enforcement. That explanation now sits at the center of the dispute: when does a concerning conversation become a credible threat that demands outside action?
The company has denied the claims. OpenAI stated that it trains models to refuse requests that could enable violence and self-harm, and that it notifies law enforcement about credible threats.
A shooting with lasting costs
On February 10, 2026, Van Rootselaar shot her mother and stepbrother at home before killing an educational assistant and five students at her former school. Van Rootselaar, who was 18 years old, died by suicide.
Nine people, primarily children, were killed in the school shooting. The listed victims included six schoolchildren and two adults, while the account of the attack also identifies an educational assistant and five students among those killed.
British Columbia says it has been forced to absorb the costs of a replacement school, mental health care, and other effects of the shooting. The province’s claim seeks financial damages tied to those costs, as well as changes to OpenAI’s handling of ChatGPT conversations that could lead to violence.
Niki Sharma, the Attorney General of British Columbia, framed the case as a question of accountability when technology intersects with criminal danger. “In any other circumstance where a person aids, encourages or conspires to commit a criminal offense or is criminally negligent, they can be investigated, prosecuted and judged,” Sharma said.
She also described reforms promised by Altman after the attack as “necessary, and yet grossly insufficient.” The lawsuit alleges that Altman promised reforms but did not follow through.
Pressure on AI safety systems
The British Columbia case adds to a growing legal challenge over the safeguards surrounding ChatGPT. More than 30 family members and others affected by the shooting have filed lawsuits in California against OpenAI.
Florida also sued OpenAI in June over safety risks, claiming ChatGPT harmed children and provided guidance on violence and self-harm. Together, these cases focus attention on the gap between detecting troubling activity and deciding what action should follow.
OpenAI spokesperson Drew Pusateri said, “OpenAI remains committed to working collaboratively with government and law enforcement officials, and continuing to advance our ongoing safety work.” The statement points toward cooperation, but British Columbia’s lawsuit demands answers about what happened before that cooperation could begin.
The case now places a sharp question before OpenAI, Altman, and the wider AI industry: what responsibility does a company carry when its safety systems identify conversations connected to violence, but its leaders decide not to contact police?
British Columbia is asking the court to address that question through damages and enforceable changes to how dangerous conversations are handled. The outcome could shape the rules companies use when AI systems flag threats, accounts are deactivated, and users return through new accounts.
The lawsuit does not erase the losses suffered by families, students, educators, or the province. It does push the technology industry toward a future where detecting risk is only the first step, and where the decision to act receives the same scrutiny as the system that raised the warning.
Based on
