BC sues OpenAI over 9 deaths in Tumbler Ridge school shooting

For legal professionals, this case tests whether AI companies can be held liable for failing to report credible threats of violence. The lawsuit alleges OpenAI's safety team flagged the shooter's gun violence chats but did not alert police, framing a potential duty to warn. It may establish California and Ninth Circuit precedent on negligence, platform immunity, and AI oversight.

On Monday, September 21, 2026, the government of British Columbia filed a lawsuit against OpenAI and its chief executive, Sam Altman, in U.S. District Court for the Northern District of California. The complaint alleges that the February 10, 2026 mass shooting at a school in Tumbler Ridge, British Columbia—in which nine people, primarily children, were killed—could have been prevented had OpenAI alerted local law enforcement after its safety team flagged the shooter's ChatGPT conversations about gun violence. The province seeks damages to cover the costs of recovery, including replacing the demolished school and providing ongoing mental health care, and an injunction requiring OpenAI to change how it handles conversations that may lead to violence. Because the case is at the pleading stage, the allegations are claims, not established findings.

“

On Monday, September 21, 2026, the government of British Columbia filed a lawsuit against OpenAI and its chief executive, Sam Altman, in U.S.

The lawsuit represents a significant escalation from individual tort claims to government-led public-cost recovery and prospective operational reform. More than 30 family members of victims and others impacted by the attack had already filed lawsuits in California federal court, but British Columbia's filing shifts the posture from private damages to sovereign cost recovery and demands for structural change. The province's attorney general had reportedly attempted to engage OpenAI after the shooting, and the complaint alleges that while Altman published a community letter saying he was "deeply sorry" OpenAI had not contacted law enforcement and promised reforms, those reforms never materialized. That allegation frames the company's post-incident statements as evidence of a broken safety commitment rather than a one-off oversight.

OpenAI's public response has been measured. Spokesperson Drew Pusateri described the shooting as an unspeakable tragedy and said OpenAI remains committed to working collaboratively with government and law enforcement officials while advancing its ongoing safety work. Notably, the statement does not directly address the specific allegation that a safety team internally flagged the shooter's conversations before the attack. The gap between a safety system alert and a police report lies at the heart of the legal claim. If discovery shows OpenAI had internal policies for escalating credible threats, the failure to follow them could support a negligence theory. If no such policies existed, the case may pivot to whether a reasonable AI company should have had them.

Jurisdictionally, the choice of a California federal court is consequential. OpenAI is headquartered in California, and U.S. courts, particularly the Ninth Circuit, have been developing early precedent on platform liability, Section 230 immunities, and the duty of online platforms to act on credible threats of physical harm. British Columbia's complaint seeks both retrospective damages and forward-looking injunctive relief, meaning the court could be asked to define the scope of an AI developer's duty to warn. That could extend far beyond this one shooting, influencing how all frontier AI companies handle violence-risk content. A ruling that creates a duty to notify police when a safety classifier flags credible threats would impose new cross-border compliance obligations on AI companies with Canadian users.

The economic stakes are substantial but difficult to quantify from the complaint alone. The province says it has been forced to absorb the costs of a replacement school, mental health care and other effects, and the school where the shooting took place is being demolished. Government lawsuits of this type bundle many categories of public expenditure—capital costs, crisis response, ongoing psychiatric services, community rehabilitation—that may not be fully covered by private insurance or existing victim compensation funds. A favorable ruling could set a template for other jurisdictions to recover similar costs from AI developers, creating a parallel track to the private family lawsuits and potentially accelerating settlement pressure.