Thirty new lawsuits filed in a California federal court against OpenAI and its CEO Sam Altman have fundamentally shifted the conversation around artificial intelligence liability. These aren't abstract policy debates about copyright or job displacement. Filed by survivors, teachers, and family members of the victims of the February mass shooting in Tumbler Ridge, British Columbia, the legal complaints cut straight to a terrifying question. What happens when an artificial intelligence company knows a user is a credible threat of violence, bans their account, but refuses to call the police?
The legal action, coordinated by the Vancouver law firm Rice Parsons Leoni and Elliott LLP alongside American co-counsel Edelson PC, brings the total number of active cases stemming from the tragedy to thirty-seven. The core allegation is straightforward and damning: internal safety teams at OpenAI flagged the shooter's account eight months before the attack, warning executives that the user posed an imminent threat of gun violence. Leadership allegedly overruled them.
Inside the Court Filings and the Warning Signs
If you're wondering how a chatbot conversation translates into a physical tragedy, the court documents lay out a harrowing timeline. The shooter, 18-year-old Jesse Van Rootselaar, used ChatGPT to plan the attack, according to the legal filings. OpenAI's own safety professionals caught the alarming activity and formally recommended contacting the Royal Canadian Mounted Police.
Instead of picking up the phone, executives chose silence. Why? The plaintiffs argue that a public scandal involving a trillion-dollar valuation and intense regulatory scrutiny motivated leadership to bury the warning. When OpenAI banned the initial account in June 2025, they stopped short of alerting law enforcement. The shooter simply bypassed the restriction by opening a second account.
Months later, the violence spilled out of a home in the remote Canadian town and into Tumbler Ridge Secondary School, leaving eight people dead—including six children—and dozens more physically and psychologically scarred.
The newest batch of lawsuits features harrowing accounts from survivors. We're talking about a 19-year-old student shot in the chest, a gym teacher who barricaded children in a conference room while listening to gunfire echo down the hall, and a student who survived by playing dead among his classmates. These personal narratives tear away the sanitized tech-industry jargon that usually surrounds AI governance.
The Defense and the Corporate Fallout
OpenAI's leadership has pushed back hard against these characterizations. Jason Kwon, OpenAI's chief strategy officer, has spoken publicly about the emotional toll the tragedy takes on the safety teams who monitor these systems daily. The company maintains that its workers care deeply about public safety and that privacy standards govern when user data can be handed over to police.
OpenAI also took aim at specific allegations in the lawsuits pointing to Chris Lehane, the company's head of global affairs and a veteran political strategist. Kwon called claims that Lehane intervened in the safety referral process "absolutely false."
Back in April, following the initial wave of seven lawsuits filed by victims' families, Sam Altman issued a formal letter of apology to the Tumbler Ridge community, admitting the company failed to notify Canadian police. Yet, an apology doesn't halt a massive legal storm or stop courts from probing how tech giants handle imminent danger signals.
Meanwhile, the local fallout in British Columbia has been profound. The provincial government has retained legal counsel in California to explore its own options for suing OpenAI. The community of Tumbler Ridge has already begun demolishing the secondary school building, desperate to turn a page on a chapter of history written in blood and gunpowder.
What This Means for the Future of Tech Accountability
We are entering a brand-new era of liability for software developers. For years, tech platforms hid behind section 230-style protections and the assumption that code is passive. ChatGPT and similar large language models complicate that defense entirely. When a system actively interacts with a user, refines ideas, and flags internal safety thresholds, it stops being a blank slate.
The lawyers driving these cases want one thing above all else: public transparency. They are pushing for the complete release of the shooter's chat logs with ChatGPT, records that OpenAI has fiercely guarded. If those logs show a clear, actionable blueprint of mass violence that was ignored for the sake of public relations, the legal foundation of the entire generative AI industry will crack wide open.
Tech companies can no longer treat real-world safety as an internal compliance metric to be managed away behind closed doors. When algorithms cross the line from text generators into predictive behavioral influences, the people building them inherit a duty of care.
If you build tools capable of recognizing human danger, you don't get to look the other way when the warning lights flash red.
