AI Company Saw Warning Signs Before Shooting, Chose Not to Alert Police

OpenAI staff spotted violent threats from one user's account in June 2025 and asked leadership to notify police. Executives refused. Eight months later, that person killed nine people at a British Columbia school. The province is suing OpenAI, saying the company's own system worked—it caught the danger—but leadership blocked the alert. The case could reshape how the law holds AI companies responsible.