British Columbia is taking OpenAI to court. The Canadian province filed a lawsuit over the company’s alleged failure to alert authorities about conversations between its chatbot and Jesse Van Rootselaar, the 18-year-old shooter in the Tumbler Ridge mass shooting.
The shooting occurred on February 10, 2026, killing six schoolchildren and two adults and injuring more than two dozen people. The lawsuit claims the attack could have been prevented if OpenAI had followed advice from its human reviewers and notified the Royal Canadian Mounted Police about Van Rootselaar’s activities on its service.
Niki Sharma, attorney general of British Columbia, announced the legal action on September 21, 2026. The announcement came after the province sued OpenAI in California, accusing the company of designing an unsafe product and acting negligently by failing to alert law enforcement about the chats.
The warning OpenAI did not act on
OpenAI’s safety team flagged the shooter’s account in 2025 because it contained potential warnings of committing real-world violence. The company banned the original account but later discovered that Van Rootselaar had created a second account.
In February, the Canadian government demanded safety changes from OpenAI after reports showed that the company had not alerted authorities when employees flagged the account. OpenAI then committed to notifying authorities about credible threats, while chief executive Sam Altman apologized for not alerting them when the first account was banned.
That apology did not satisfy British Columbia premier David Eby. He called it “necessary, and yet grossly insufficient.” The lawsuit claims OpenAI has not meaningfully engaged with Canada and British Columbia on safety matters since the shooting.
The legal action also seeks payment for construction of a new school in Tumbler Ridge. A 13-year-old identified as “A.C.” is among the plaintiffs, and the victims’ families have filed a separate complaint accusing OpenAI of negligence.
A broader test for AI accountability
Sharma framed the case as a question of whether AI companies should face the same scrutiny applied when other products or services contribute to criminal harm. “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,” she said.
She also said AI companies’ CEOs know their technologies pose a safety risk, pointing to calls from some of them for a global slowdown of frontier AI development. That argument places OpenAI’s internal warnings at the center of the dispute: the company’s own reviewers allegedly identified a danger, yet authorities were not notified.
The lawsuit does not merely challenge one missed alert. It accuses OpenAI of building a product whose safety systems failed at the moment they mattered, then failing to work meaningfully with Canadian and British Columbian officials afterward.
OpenAI’s commitments after the shooting now face a legal test. The company says it will notify authorities about credible threats, but British Columbia’s lawsuit asks what responsibility existed before that promise—and what consequences follow when internal safety warnings do not reach law enforcement.
The case also adds pressure from multiple directions: the provincial government is suing in California, the victims’ families have filed their own complaint, and the province wants a new school built in Tumbler Ridge. AI safety has spent years living in policy statements and conference panels. This time, it has arrived in court.
Based on
