B.C. Sues OpenAI Over Tumbler Ridge — ChatGPT Flagged, RCMP Not Told

33

On Monday, Sept. 21, 2026, British Columbia filed suit against OpenAI in California federal court over the company’s alleged failure to warn law enforcement about ChatGPT platform threats before the Feb. 10 Tumbler Ridge Secondary tragedy. Attorney General Niki Sharma announced the action with School District 59 Peace River South as co-plaintiff — coverage from the B.C. government news release, the Times Colonist, and the Globe and Mail.

This is a provincial government plaintiff stepping into U.S. product-liability and negligence territory against a frontier chatbot lab. It is separate from the survivor and family cases already on file.

What The Province Alleges

Per Sharma’s statement and reporting in the Times Colonist and Globe and Mail, the complaint alleges negligence, product defect, and “aiding and abetting.” The province says OpenAI confirmed that the shooter’s ChatGPT conversation was alarming enough to trigger an internal employee review — was not reported to the RCMP — and was later dismissed. Sharma says the province asked for the chats and that OpenAI refused.

Damages sought, as reported, include costs tied to replacing the secondary school and other extraordinary public outlays, plus injunctive relief aimed at stronger conversational safeguards. OpenAI CEO Sam Altman is named as an individual defendant alongside the company, according to those outlets.

These are allegations in a filed complaint, not adjudicated findings. Nothing in this piece treats the duty-to-warn theory, product-defect claims, or causation arguments as proven fact.

Sharma’s government statement frames the suit as accountability for Tumbler Ridge and as a broader push for national AI safeguards — including Criminal Code amendments she says she has urged on federal ministers. The province argues technology companies that learn of credible violence threats carry responsibilities that go beyond an internal file note. Those arguments remain contested in court.

OpenAI’s Response

In a statement to Business Insider, an OpenAI spokesperson called the shooting “an unspeakable tragedy” and said the company “remains committed to working collaboratively with government and law enforcement officials, and continuing to advance our ongoing safety work.”

BBC coverage notes OpenAI previously apologized for not flagging the account to law enforcement; Altman has said he was “deeply sorry” the company did not alert authorities. The provincial filing is distinct from those earlier expressions of regret — and from the more than 30 survivor and family suits Business Insider reports were filed by early September.

JamoraquAI Take

When a lab’s own safety stack flags credible violence risk, “we reviewed it internally” is not the same as a duty to warn — and a provincial government filing in California turns that gap into a product-liability stress test for chatbots worldwide.

Internal review without external escalation is a process choice. A government plaintiff is asking a California court to treat that choice as a legal failure when credible violence risk was already on the company’s radar. Whether courts agree is for litigation. The industry question lands now: if your own systems raise a hand, what is the minimum public-safety bar before “we looked at it” counts as enough?

For chatbot makers everywhere, the Tumbler Ridge filing reframes threat reporting from a soft ops judgment into a hard product-liability stress test — with a sovereign plaintiff, injunctive asks, and a named CEO in the caption.

Sources

LEAVE A REPLY

Please enter your comment!
Please enter your name here