British Columbia is suing OpenAI, the company behind ChatGPT, because the company's safety team flagged violent requests from a teenager before he killed nine people in a February 2026 school attack Al Jazeera.

Here's what happened: In June 2025, OpenAI's safety system detected that an 18-year-old named Jesse Van Rootselaar was using ChatGPT to ask violent questions and plan attacks. OpenAI deleted his account. But the company did not tell police. Eight months later, in February 2026, Van Rootselaar killed his mother and half-brother, then attacked Tumbler Ridge Secondary School in a remote area of northeastern British Columbia. He killed five children and one teacher before taking his own life. Twenty-seven others were wounded.

According to the province's lawsuit, twelve OpenAI employees told the company to contact the police about what they had found Al Jazeera. The company said no. OpenAI argued that Van Rootselaar's activity did not reach what they called "an imminent and credible risk"—their own standard for alerting authorities Reuters.

Victims' families have also filed their own lawsuit in California court against OpenAI and CEO Sam Altman Reuters. Their lawyers argue that once OpenAI's safety team spotted the danger, the company had a responsibility to warn authorities The Guardian. OpenAI says Van Rootselaar had mental health problems, and that those—not the ChatGPT prompts—were the real cause Reuters.

In April, Sam Altman published a public apology in the town newspaper saying he deeply regretted OpenAI's failure to alert the police Tumbler RidgeLines. Yet OpenAI continues fighting the lawsuits in court—even as the CEO has admitted the company made a mistake.

Why is this case important? Because it raises a question no courtroom has fully answered before: When an AI company's own safety team detects a threat, does the company have a legal duty to call the police? The answer could reshape how AI companies operate across North America.

Canada's government has already started moving. After the shooting, federal officials brought OpenAI's safety team to Ottawa to demand better protections Reuters. Disappointed with the company's response, the government told OpenAI to strengthen its safeguards or face new laws Reuters. In June, Ottawa passed legislation to regulate chatbots, directly tied to what happened in Tumbler Ridge—though industry opposition means the law may face challenges Reuters.