British Columbia Accuses OpenAI of Failing to Prevent Tumbler Ridge Violence

British Columbia argues that ChatGPT’s alleged failures were rooted in OpenAI’s design choices—not in the actions of its users. The province claims OpenAI disabled Tumbler Ridge’s ability to end chats and later introduced a more flattering model that the company already knew could assist users attempting to cause real-world violence.

Since ChatGPT’s release, its model specifications have instructed chatbots to assume “good faith” and “no probing intent” when users initiate conversations about violence. British Columbia is asking the court to order OpenAI to change those instructions.

Province Seeks ChatGPT Logs in Tumbler Ridge Case

The Tumbler Ridge community wants access to the chat logs, as does the province of British Columbia, which is funding the recovery effort.

The province argues that the logs could show ChatGPT dangerously “furthered the shooter’s mental instability” by encouraging, elaborating on, and reinforcing violent thoughts “instead of interrupting or directing users to real-world help.”

If OpenAI loses the case, it could face a broad range of damages, including an order to reimburse emergency response costs. Punitive damages may also be awarded if the court determines that the alleged failure to warn police was intentional.

According to the complaint, OpenAI “repeatedly reiterated to its users and the public that it detects users who pose a risk of harm to others, routes such conversations to authorized and trained reviewers, and reports imminent threats of serious physical harm to law enforcement.”

“OpenAI cannot waive its publicly assumed obligation, which it has voluntarily assumed and claimed to be fulfilling the very protective function required by the warning duty law,” the complaint states.

British Columbia Says Sam Altman’s Apology Is Not Enough

In April 2026, OpenAI CEO Sam Altman publicly acknowledged that the company had failed the Tumbler Ridge community.

“We deeply regret that we did not alert law enforcement about the account that was banned in June,” Altman said.

British Columbia disputes that account of events, arguing that the account was never banned. The province describes Altman’s statement as an “admission” that OpenAI “failed to notify law enforcement after our safety team identified the risk of violence.”

Province Calls for Independent OpenAI Safety Audits

To ensure that the most violent conversations are consistently reported to police, British Columbia has asked the court to require OpenAI to conduct regular independent audits.

The complaint describes the company’s conduct as “despicable.” It also argues that “meaningful user bans” could have prevented the shooter from continuing to access ChatGPT.

According to the lawsuit, continued access to the chatbot “made the shooter more lethal.” The suit further alleges that, if OpenAI had reported the user to law enforcement, the gunman could have already been on the radar of local authorities and the weapon used in the shooting might have been removed from the home.