British Columbia sues OpenAI over school shooting, alleging preventable tragedy

British Columbia filed suit against OpenAI and CEO Sam Altman in San Francisco federal court on Monday, claiming the company could have prevented a school shooting that killed nine people if it had simply told police what it knew. The lawsuit centers on a straightforward allegation: OpenAI's own safety team flagged violent conversations on ChatGPT in June 2025, recommended alerting law enforcement, and was overruled by leadership. The shooter, Jesse Van Rootselaar, then obtained a new account after her original one was deactivated, and used the platform to continue planning the attack that would unfold on February 10.

On that day, Van Rootselaar, who was 18, shot her mother and stepbrother at home. She then went to her former school in Tumbler Ridge and killed an educational assistant and five students, all aged 12 or 13. She died by suicide at the scene. The toll was nine dead, most of them children. The school building itself is now being demolished. British Columbia is absorbing the costs of a replacement facility, mental health services, and the cascading expenses of recovery—all of which the province is now asking OpenAI to pay for through damages.

The lawsuit also seeks a court order forcing OpenAI to change how it handles conversations that suggest potential violence. This is not a theoretical concern. According to the lawsuit, citing OpenAI whistleblowers who spoke to the Wall Street Journal, the company's safety systems had identified Van Rootselaar's misuse of the model "in furtherance of violent activities." The internal bar for reporting to law enforcement, however, was apparently higher than what the safety team believed the situation warranted. OpenAI later said the flagged conversations did not meet its criteria for alerting police.

Sam Altman published a letter to the community after the shooting saying he was "deeply sorry" the company had not contacted law enforcement. He promised reforms. According to the lawsuit, those reforms never materialized, despite efforts from British Columbia's attorney general to engage with the company on the matter. OpenAI's current statement, delivered through spokesperson Drew Pusateri, calls the shooting "an unspeakable tragedy" and reaffirms the company's commitment to safety work and collaboration with government and law enforcement. The company maintains that it trains its models to refuse requests that could "meaningfully enable violence" and notifies law enforcement when conversations suggest "an imminent and credible risk of harm to others."

British Columbia's lawsuit is one of a growing wave of legal actions against OpenAI over ChatGPT's role in facilitating violence. More than 30 family members of victims and others affected by the Tumbler Ridge shooting have already filed their own lawsuits in California. In June, Florida became the first state to sue OpenAI, accusing the company of misrepresenting the safety of its platform and claiming ChatGPT had provided information to school shooters and guidance on self-harm. Other cases have been filed in recent months related to a mass shooting at Florida State University. OpenAI has denied all such claims.

The core tension in the Tumbler Ridge case is whether a company that detects dangerous behavior on its platform has an obligation to report it to authorities, and at what threshold that obligation kicks in. OpenAI's position is that its safety systems worked as designed—they flagged the problem, the company took action by deactivating the account, and the conversations did not cross the internal line for law enforcement notification. British Columbia's position, backed by its own whistleblower evidence, is that the company had enough information to prevent a tragedy and chose not to act. The lawsuit will test whether courts agree.