British Columbia says OpenAI could have stopped a school shooting, and is suing over it

midian182

Posts: 11,938   +183
Staff member
What just happened? OpenAI and Sam Altman are being sued by British Columbia over claims that the company could have prevented a mass shooting in the Canadian province. The lawsuit alleges that OpenAI failed to inform authorities about the shooter's violent ChatGPT conversations.

The suit, which was filed in San Francisco federal court, accuses OpenAI accuses OpenAI of designing an unsafe product and acting negligently. It seeks damages to fund recovery efforts in the province following the shooting at a school in Tumbler Ridge in February.

The British Columbia government seeks financial compensation from the tech giant for past and future expenses related to emergency response and community recovery. It also asks the court to require the company to change how it detects and responds to user conversations involving threats of violence.

Nine people, most of whom were children, were killed in the school shooting. The shooter killed her mother and half-brother at their home before going to the secondary school, where she killed five children aged 12 and 13 and an educational assistant. She then died by suicide.

The shooter, eighteen-year-old former pupil Jesse Van ⁠Rootselaar, had previously been flagged by OpenAI's safety team for her conversations about gun violence but OpenAI never alerted police because the conversations did not meet the threshold for legal referral.

The suit claims that the shooting would have been averted if the company had followed its human reviewers' advice to notify the Royal Canadian Mounted Police about Rootselaar's activities.

Altman apologized in April for OpenAI's failure to contact police at the time. The lawsuit says that while he promised reforms, the company never ‌followed through, despite efforts from British Columbia's attorney general to engage.

The school where the shooting occurred is now being demolished, while the province has had to cover the cost of a replacement school, mental health support, and other expenses stemming from the attack.

Van Rootselaar's actions were first flagged by OpenAI in June 2025. The lawsuit alleges that members of OpenAI's safety team urged the company to contact police, but Altman and other senior executives rejected the recommendation. It cites OpenAI whistleblowers who spoke to The Wall Street Journal after the shooting. Her account was deactivated, but she opened a new account and continued using the AI to plan her attack.

This isn't the first time OpenAI has faced legal action connected to a mass shooting. In June, Florida sued OpenAI and Altman, alleging that ChatGPT had contributed to real-world harms, including the 2025 Florida State University shooting. The lawsuit followed a criminal investigation launched by Florida's attorney general into OpenAI and ChatGPT over the incident. That probe was later expanded after prosecutors alleged that a suspect in the murders of two University of South Florida students had also used ChatGPT before the killings.

Permalink to story:

 
Yeah, no.

The moment one AI system misunderstands or misinterprets what I’m saying, and/or if I’m just being a smartass - and the police are called… that’s the moment I stop using the tool.

We don’t need third party AI models, who aren’t even being effectively controlled by their own company to have the ability to contact LE autonomously.
 
In case you're looking for a textbook definition of nonsense, look no further - you've found it.

Search engines and chatbots are being queried billions of times daily for all sorts of stuff. If they start reporting every vaguely 'suspicious' query to the authorities, this will simply overload them and drastically reduce their efficiency in handling actual threats.
People going crazy is totally unpredictable, and has always been unpredictable.

 
Open AI should turn around and sue the school and the parents for not realizing their student/child was in mental distress. Everyone wants to duck their responsibilities and point the finger elsewhere. If Open AI were to report everyone that has used their services to contemplate a crime they would get sued six ways to Sunday.
 
Yeah, no.

The moment one AI system misunderstands or misinterprets what I’m saying, and/or if I’m just being a smartass - and the police are called… that’s the moment I stop using the tool.

We don’t need third party AI models, who aren’t even being effectively controlled by their own company to have the ability to contact LE autonomously.
Right, AI isn't smart enough to determine actual intent or that the person has the means to commit a crime. And due to AI hallucinations I could see a lot of people who did absolutely nothing getting caught up in a fishing expedition.
 
Back