Lawsuit: OpenAI hid mass shooter’s chat logs to avoid outing other violent users

OpenAI's handling of violent chat logs could lead to new legal obligations to report threats and reveal user data.

British Columbia is suing OpenAI, seeking financial compensation and injunctive relief after a shooter used ChatGPT to plan a deadly attack. The province alleges that OpenAI failed to warn law enforcement about the shooter, Jesse Van Rootselaar, who killed eight people in February 2026 before dying by suicide at Tumbler Ridge Secondary School. The lawsuit claims OpenAI hid the shooter's chat logs, which human reviewers had flagged as a credible threat, to avoid outing other violent users and protect its potential IPO. The province argues that OpenAI should fund the demolition of the school and the construction of a new one, along with other emergency and long-term recovery costs, including extensive mental health services for the traumatized community. British Columbia is also demanding OpenAI update ChatGPT to automatically terminate violent conversations.

This lawsuit raises critical questions about AI developer responsibility and the extent to which they can be held liable for harms facilitated by their products. The demand for OpenAI to proactively terminate violent conversations suggests a potential shift towards AI systems having built-in mechanisms to identify and interrupt dangerous user intent, moving beyond simple content moderation.

The case also highlights the tension between user privacy and public safety, particularly when AI tools can be used to plan real-world violence. OpenAI's alleged decision to withhold logs to protect its IPO could set a precedent for how technology companies balance profit motives with ethical obligations and regulatory oversight in the face of catastrophic events.