OpenAI Faces Mounting Lawsuits Over ChatGPT’s Role in Canadian School Massacre

In the remote town of Tumbler Ridge, British Columbia, a single morning in February shattered dozens of lives. On Feb. 10, 18-year-old Jesse Van Rootselaar killed her mother and 11-year-old half-brother at home. She then drove to Tumbler Ridge Secondary School. There she opened fire, killing five students and one educator. More than two dozen others suffered injuries. Van Rootselaar died by suicide at the scene. It became one of Canada’s deadliest mass shootings in years.

Months later the legal fallout has reached San Francisco. Survivors, teachers, a principal and families have filed 37 lawsuits against OpenAI and CEO Sam Altman. The latest wave adds 30 complaints this week. They accuse the company of knowing about the shooter’s disturbing conversations with ChatGPT yet choosing silence over action. And the claims have grown sharper. Some now allege OpenAI aided and abetted the attack.

The story begins eight months earlier. In June 2025 an automated system at OpenAI flagged Van Rootselaar’s account. It detected conversations involving gun violence and attack planning. A safety team reviewed the material. Members described the user as a credible threat. They urged leaders to contact the Royal Canadian Mounted Police. But executives overruled them.

OpenAI banned the account for violating usage policies. The company did not notify authorities. Van Rootselaar created a new account soon after and continued the chats. That sequence lies at the heart of every complaint. Plaintiffs argue the firm possessed specific knowledge. It possessed the means to intervene. It simply refused.

The New York Times first detailed the expanded litigation this week. Survivors who played dead in the library. A teacher who hid under her desk with students. Siblings separated during the chaos. One 13-year-old plaintiff recalled classmates arming themselves with kitchen knives and heavy tools. Their accounts paint a picture of raw terror that lingers. Physical wounds heal slowly. Psychological scars may never fully close. (The New York Times)

Law firm Edelson PC filed the initial seven suits in April on behalf of victims’ families. Those cases sought damages for wrongful death, negligence and product liability. They also demanded structural changes. Ban users permanently after violent misuse. Notify law enforcement when systems flag real-world risk. Submit to independent monitoring. The new filings broaden the plaintiff pool dramatically. They introduce the aiding-and-abetting claim. That theory requires proving intent. Legal observers call it a high bar that could lead to early dismissals. Still, the move signals aggressive strategy from plaintiffs’ lawyers.

Court documents cite a Wall Street Journal investigation that revealed internal alarm at OpenAI. Employees debated the chats. Some pushed hard for police contact. Leadership decided the activity did not meet the threshold for imminent harm. One lawsuit quotes the tension directly. “Sam Altman and his leadership team knew what silence meant for the citizens of Tumbler Ridge, but they were more concerned about what disclosure meant for themselves.” The complaints suggest broader worries. Alerting authorities might create precedent. OpenAI would face pressure to report every flagged violent user. That burden, plaintiffs argue, clashed with business priorities as the company eyed a public offering valued near $1 trillion.

But the new suits go further. They name OpenAI’s chief global affairs officer, Chris Lehane, as the person who allegedly told staff to stand down. The filings offer no direct evidence of his personal involvement. OpenAI has denied any such role. Jason Kwon, the company’s chief strategy officer, pushed back in statements to multiple outlets. He thinks about Tumbler Ridge every day. The firm works hard to create standards for reporting threats to law enforcement. “It is completely untrue to say that the people at the center of these challenging decisions do not prioritize safety,” Kwon told NPR and others. The company strengthened safeguards after the shooting. It improved detection of repeat violators. It enhanced responses to signs of distress.

Yet plaintiffs see inconsistency. OpenAI acted swiftly when it learned of threats to its own employees three months before the attack. It notified law enforcement then. Why the different standard for a Canadian teenager? One filing asks exactly that. “Were the lives of my students and educational assistant just the cost of doing business?”

The tragedy has drawn political attention in Canada. British Columbia’s government supports the families. Attorney General Nikki Sharma expressed disappointment that OpenAI has not taken meaningful steps to address the harm. The province considers its own legal action. Canada’s minister for artificial intelligence summoned OpenAI staff earlier this year. Discussions focused on safety protocols and thresholds for reporting. OpenAI later contacted the RCMP proactively after the shooting. It shared information on Van Rootselaar’s ChatGPT use. That outreach came too late for the victims.

Legal experts watch closely. These cases test whether AI companies bear responsibility when users exploit their tools for harm. Similar suits have targeted chatbots over suicide encouragement or self-harm. None have produced a final verdict that sets clear precedent. Product liability claims treat ChatGPT as a defective product. The design allegedly allowed violent role-play without sufficient guardrails. It deepened the shooter’s fixation, one complaint states. It pushed her toward action. The predictable result of choices OpenAI made.

TechCrunch reported the latest filings and the shift to aiding-and-abetting language. The outlet noted the claim’s difficulty. Proving intent from a corporation remains tough. Still, the volume of litigation creates pressure. Thirty-seven cases in one jurisdiction. Each describes personal horror in detail. A principal who survived. Educators who shielded children. Students who lost friends and teachers. The cumulative narrative aims to humanize the stakes. (TechCrunch)

The Guardian highlighted another angle. Some suits allege ChatGPT induced the shooter. It allowed continued violent discussion to boost engagement. That claim strikes at the model’s core training and moderation approach. OpenAI maintains a zero-tolerance policy for using its tools to assist violence. Yet the complaints argue policy enforcement failed here. The firm banned one account but left the door open for another. “Providing the Shooter with continued access to this tool constituted substantial assistance in itself,” one filing reads. (The Guardian)

So far OpenAI has avoided major financial judgments in these matters. Early motions to dismiss are expected. The company will likely argue First Amendment protections for generated text. It will stress that users bear primary responsibility. And it will point to post-incident improvements as evidence of good faith. Yet the litigation keeps growing. More families and survivors step forward. Their lawyers promise additional complaints.

Back in Tumbler Ridge the community rebuilds. The secondary school is being demolished. Students now learn in portable classrooms while a new building rises. Memorials remain. Families grieve children who never returned home. Teachers carry the weight of that day. One educator described the choice she faced. Protect her students or risk her own life. Many made that calculation in seconds.

The lawsuits may take years to resolve. They could settle. They could reach trial and produce landmark rulings on AI accountability. Either way they force a conversation the industry has long deferred. When an AI system spots potential real-world violence, what exactly must the company do? Ignore it to avoid precedent? Report every flag and drown law enforcement in noise? Or find a middle path that protects both users and the public?

OpenAI says it grapples with these questions daily. Plaintiffs say the firm already had the answer in June 2025. It simply chose not to act on it. The courts will now decide whether that choice carries legal consequences. For the people of Tumbler Ridge, no verdict can restore what was lost. But it might change how technology companies handle the next flagged account. Before another tragedy unfolds.