British Columbia Sues OpenAI Claiming ChatGPT Planning Warnings Could Have Stopped a School Massacre

British Columbia has sued OpenAI and its chief executive, Sam Altman, arguing that a February school shooting that killed nine people could have been prevented if the company had told police the gunman used ChatGPT to plan it. The province filed the case in San Francisco federal court on Monday.

The attack hit Tumbler Ridge, a small mining town in northern British Columbia. Jesse Van Rootselaar, 18, shot her mother and stepbrother at home, then killed an educational assistant and five students aged 12 and 13 at her former school on 10 February, according to police. She then died by suicide. It ranks among the worst mass shootings in Canadian history.

The suit, brought by Attorney General Niki Sharma, seeks damages to fund recovery in the province and a court order changing how OpenAI handles ChatGPT conversations that could lead to violence. "The people of Tumbler Ridge have endured an unimaginable loss, and they deserve answers and accountability," Sharma said.

The central allegation is about a warning that never reached police. OpenAI's safety team first flagged Van Rootselaar's conversations about gun violence in June 2025 and recommended contacting law enforcement, but Altman and other leaders overruled them, the lawsuit says. It bases that account on OpenAI whistleblowers who spoke to the Wall Street Journal after the shooting.

Her account was deactivated. She opened a new one and kept planning the attack.

After the Journal's report, OpenAI said the account had been flagged by systems that identify misuses of its models in furtherance of violent activities, but that the case did not meet its internal criteria for reporting to law enforcement.

Altman published a letter to Tumbler Ridge in April saying he was "deeply sorry" the company had not contacted police, and promised reforms. The lawsuit says those changes never materialized, even after Sharma's office tried to engage.

The costs are still mounting. The school where the children died is being demolished, and the province says it has been forced to cover a replacement, along with mental health care and other fallout. One survivor, 13-year-old Maya Gebala, recovered after being shot three times, including in the skull, and now lives with permanent brain injuries.

OpenAI spokesperson Drew Pusateri called the shooting an unspeakable tragedy and said the company remains committed to working with government and law enforcement. OpenAI has denied similar claims elsewhere, saying it trains its models to refuse requests that could meaningfully enable violence and alerts police when a conversation suggests an imminent and credible risk of harm to others.

The province is not alone. More than 30 family members of victims have sued OpenAI in California, and the family of a victim in a Florida State University shooting has filed its own case. Florida in June became the first state to sue the company over safety risks. The fights sit alongside other legal pressure on the company, from a media copyright suit to scrutiny of how it processes and stores sensitive chats.

For now this is a civil complaint, not a criminal case, and no judge has tested its claims. But it makes British Columbia the first government plaintiff arguing that OpenAI's own safety flags carried a duty to warn the police who might have stopped the attack.