British Columbia sues OpenAI and Sam Altman over the Tumbler Ridge school shooting

The Canadian province of British Columbia sued OpenAI and its CEO Sam Altman in federal court in San Francisco on September 21, 2026, alleging the company could have used ChatGPT conversation logs to warn police ahead of a mass shooting at a school in Tumbler Ridge, BC.

What's new

Bloomberg reported that British Columbia "filed a lawsuit in California against OpenAI, alleging it could have used ChatGPT logs to warn police and avert a mass shooting in the Canadian province earlier this year." The suit centers on the February 10, 2026 attack, in which 18-year-old Jesse Van Rootselaar killed five students and a teacher at Tumbler Ridge Secondary School, along with her mother and half-brother, before taking her own life.

The province's claim rests on negligence: it alleges OpenAI's own safety team had flagged Van Rootselaar's conversations about gun violence months before the attack but the company did not notify law enforcement. British Columbia is seeking financial compensation for the costs of the emergency response and the province's ongoing recovery efforts, plus a court order requiring OpenAI to change how it identifies and escalates conversations that show signs of imminent violence. The suit names both OpenAI and Altman as defendants and was filed in the U.S. rather than Canada.

OpenAI has already acknowledged missing the warning signs in this case. Altman published an apology in April 2026 saying he was "deeply sorry" the company had not contacted law enforcement after flagging the account, and promised reforms to how threatening conversations get escalated. British Columbia's suit alleges those promised reforms did not materialize, despite outreach from the provincial attorney general's office.

Context

British Columbia is not the first plaintiff to sue OpenAI over this shooting. More than 30 other lawsuits — from victims' families, survivors, and others affected — have already been filed over the Tumbler Ridge attack, most of them in U.S. federal court, since families of victims first sued OpenAI and Altman in the spring. What makes this filing different is the plaintiff: a Canadian provincial government, not an individual family, is now suing an American AI company in a U.S. court over an alleged failure to act on safety signals its own systems generated. It's part of a wider pattern of litigation this year linking chatbot conversations to violent incidents, including a similar state-led suit filed by Florida over a separate case.

Why it matters

The legal theory here doesn't turn on what ChatGPT said to the shooter — it turns on what OpenAI's safety systems already knew and didn't act on. If British Columbia's account holds up, the case argues OpenAI had identified the risk internally and chose not to escalate it, which is a narrower and potentially harder claim for OpenAI to defend than the broader claims in other AI-harm lawsuits about what a chatbot generated. A government plaintiff also raises the stakes: unlike individual families, British Columbia can point to concrete public costs (emergency response, recovery) as damages, and its request for a court-ordered change to OpenAI's escalation process, if granted, would reach beyond the shooting itself into how OpenAI handles every account its safety systems flag for violence risk going forward.