Could AI firms be held responsible for mass shootings?

British Columbia has lodged a lawsuit in U.S. federal court accusing OpenAI of failing to notify authorities about troubling user interactions on ChatGPT prior to the Tumbler Ridge mass shooting. The provincial complaint, filed in San Francisco, alleges OpenAI "aided and abetted a mass shooting" and asserts product defect liability; it seeks funds to cover emergency response and community recovery costs and requests court-ordered changes to OpenAI's procedures for identifying and handling user conversations that threaten violence. The suit also names OpenAI CEO Sam Altman as a defendant. The legal actions follow earlier filings by about 30 victims and family members in the Northern District of California, who contend ChatGPT played a role in the attack and that OpenAI could and should have prevented it. OpenAI has said it deactivated a ChatGPT account tied to the attacker eight months before the shooting for misuse related to violent activities, but the company has maintained that the content did not meet its internal threshold for referral to law enforcement at that time. Reporting from U.S. outlets has added detail to the dispute over OpenAI's response. The Wall Street Journal reported that around a dozen employees debated whether to alert police after reviewing conversations in which the user described scenarios with gun violence, but that leadership decided not to refer the matter. Separately, Mother Jones reported that the banned user later opened a second account and that chat logs showed the model offering advice on how to avoid its own safety measures by framing violent content as fictional or hypothetical. OpenAI says it has provided chat logs to the Royal Canadian Mounted Police; British Columbia says it has not received those materials. The lawsuits raise broader legal and policy questions about the responsibilities of AI platform operators. Experts note that British Columbia may have brought its case in a U.S. court in part because Canadian laws limit damage awards and punitive damages, potentially reducing recovery for provincial costs. Plaintiffs seek not only compensation but also changes to how AI companies detect serious threats, enforce user-level bans and coordinate with authorities — issues that could influence industry practices and cross-border legal strategies as more cases test the limits of existing liability and duty-to-warn doctrines.