A Province Takes the Stand: British Columbia Sues OpenAI and Sam Altman Over Tumbler Ridge

The entity walking into a San Francisco federal courthouse this time is not a grieving family or a class-action firm. It is a Canadian province. On Monday, September 21, 2026, British Columbia Attorney General Niki Sharma announced in Victoria that the province — filing jointly with the Peace River South School District (SD59) — had launched a lawsuit against OpenAI over its alleged role in the February 10 Tumbler Ridge school shooting. The suit names OpenAI CEO Sam Altman as an individual defendant, and it rests on a simple, brutal theory: the company identified a credible threat on its own platform, chose not to tell the police, and a community has been paying for that choice ever since.

What the province is actually claiming

This is not another wrongful-death complaint stacked onto the pile. Sharma was explicit that the province’s claim stands apart from the 37 private lawsuits already filed by victims’ families, educators, and students in California, where OpenAI is headquartered. “Our focus is on the losses that the province suffered as a result of the conduct and harm,” she said, “so the basis for our claim for damages is quite different.”

What does a province lose when a school shooting happens? B.C. says it has spent significant resources on the recovery and rebuilding of Tumbler Ridge — educators, clinicians, victim-services workers, and police time, alongside physical infrastructure. The suit seeks reimbursement for the construction of a new secondary school; the old building was demolished in August, with the site to be restored to grass, because the community could not imagine teaching inside those walls again.

The claim also targets product design, not just disclosure failure. Sharma said the province is suing OpenAI both for its failure to notify law enforcement of threats made on ChatGPT before the shooting and for designing an unsafe product. That two-track theory — duty-to-warn plus defective design — mirrors the architecture of the private litigation, but a government plaintiff changes the political arithmetic entirely. Regulators watch what other regulators do.

The apology that became evidence

The most striking line in the filing quotes Sam Altman himself. In April, after the first wave of suits, Altman published an open letter apologizing to the Tumbler Ridge community for not notifying Canadian police about what OpenAI knew: “I am deeply sorry that we did not alert law enforcement to the account that was banned in June.” The province’s complaint treats that sentence not as contrition but as proof — describing it in the claim as an “admission that OpenAI identified the risk and failed to act on it.”

The timeline underneath is already largely public record. OpenAI’s safety team flagged the shooter’s account — 18-year-old Jesse Van Rootselaar, who killed six children, one teacher, and her own family members before dying of a self-inflicted gunshot wound — after violent queries in June 2025, roughly eight months before the attack. The company deactivated the account. Internal deliberations over whether to alert the RCMP were overruled, according to prior filings by plaintiffs’ attorney Jay Edelson, who has alleged the decision ran up through head of global affairs Chris Lehane and was sanctioned by Altman. OpenAI has previously and categorically rejected that account, with Chief Strategy Officer Jason Kwon calling the claims about Lehane “absolutely false.”

“We have asked OpenAI to disclose these chats. They have refused.”

At her Monday news conference, Sharma added a detail that will pressure OpenAI in the court of public opinion regardless of how the legal arguments land. She said the conversations between the shooter and ChatGPT were alarming enough to trigger an internal review at OpenAI — but were never reported to the RCMP. The province asked OpenAI to hand over the chat logs. The company declined.

“We have asked OpenAI to disclose these chats. They have refused. We should all be asking them: Why?” Sharma said. She noted that she has not read the conversations herself, and that the decision not to report them was made in California — which is precisely why a Canadian province filed suit in a U.S. district court rather than at home. Jurisdiction follows the decision.

OpenAI’s response to the filing was brief. In an emailed statement carried by CBC, the company called Tumbler Ridge an “unspeakable tragedy” and said its thoughts remain with the victims, their families, and the community, adding that it “remains committed to working collaboratively with government and law enforcement officials, and continuing to advance our ongoing safety work.”

Three fronts now open

The legal exposure around Tumbler Ridge now runs on three distinct tracks:

1. Private civil litigation. Thirty-seven claims from families, survivors, teachers, and students, filed in the Northern District of California, alleging negligence, product liability, and aiding and abetting a mass shooting. Discovery there is already surfacing internal deliberations.
2. The province’s suit. B.C. and SD59’s new claim, seeking damages for public costs, a rebuilt school, and — critically — “injunctive relief to ensure that this tragedy does not happen again,” language that aims at changing how the product itself works.
3. Criminal investigation. The B.C. RCMP probe remains open. In July, police said digital and physical evidence had been collected and analysts were still working through seized electronic devices, adding: “This continues to be a criminal investigation, in which we have not eliminated the possibility that charges could be laid.”

Why a government plaintiff matters

Private suits can be settled, sealed, and folded into a quarterly earnings line. A suit brought by a province cannot. Sharma framed the action as being about accountability and structural change: “The lawsuit being filed raises serious questions about the responsibilities of technology companies in ensuring their products do not cause harm. We believe that companies developing and deploying artificial intelligence must be held accountable when their products contribute to real-world harm.”

She is also pushing the pressure upstream to her own federal government. Sharma said she has written to federal ministers Fraser, Solomon, and Miller with proposals to amend Canada’s Artificial Intelligence and Data Act, urging Ottawa to set clear national AI safety standards. “This case highlights the urgent need for strong national safeguards for artificial intelligence technology,” she said.

For OpenAI, the timing could hardly be worse. The company is simultaneously navigating a White House standoff over Anthropic, an industry-wide debate over pacing frontier development, fresh disclosures of “concerning” model behavior incidents, and the aftermath of a researcher breach that used a rival’s model to write the exploit. Each of those stories is, in isolation, survivable. A provincial government with subpoena power, an explicit theory of institutional failure, and a public refusal-to-disclose narrative is a different order of problem — because it converts a tort docket into a standing public inquiry.

None of the allegations have been proven in court. But the burden of the next phase — depositions under oath, internal documents, and possibly those chat logs — now sits with a company whose CEO’s own apology is exhibit A in the complaint against him.