Tumbler Ridge shooter’s ChatGPT account banned in 2025 for references to violence

OpenAI, the maker of ChatGPT, says it banned the account of Tumbler Ridge shooting suspect Jesse Van Rootselaar in 2025 after it was flagged internally as being misused in “furtherance of violent activities.”

In a statement to 1130 NewsRadio, the company says it considered referring the account to law enforcement, but determined the activity didn’t meet the threshold to do so.

In order to meet that threshold, the activity would have had to indicate “an imminent and credible risk of serious physical harm to others,” which it did not appear to do, the company says.

After hearing of the mass shooting in Tumbler Ridge, OpenAI says it “proactively” reached out to the RCMP with information on the suspect’s use of ChatGPT and says it will continue to support the investigation. The B.C. government says OpenAI had a preplanned meeting with provincial officials the day after the shooting, but the company made no mention of concerns it had regarding the shooter’s online interactions with its ChatGPT program. The province says the day after the meeting, OpenAI requested contact information for the RCMP.

OpenAI says it avoids over-enforcement of these policies because it can be distressing when, for example, police show up at the account holder’s home unannounced. This also raises privacy concerns, it says.

Sharon Bauer, an expert on digital privacy, says disclosure of personal information without consent is a very rare circumstance in Canada, but information that reflects an imminent risk of serious bodily harm is an exception. It is hard to find the right balance, she says.

“There is this fine line between protecting someone’s privacy and protecting the safety and security of individuals,” Bauer said.

“It appears there was some human error, of course in hindsight we know that.”

Referrals to law enforcement are just one of the company’s approaches to mitigating risks, OpenAI says.

This is not the only example of alarming online activity by the shooter. Van Rootselaar’s Roblox account was also banned after the game developers found it was being used in a game encouraging users to go on a virtual shooting spree. As well, the suspect also made several posts detailing psychotic breaks, childhood traumas, and an interest in mass shooters.

And, as first reported by the New York Times, posts attributed to her on a different website dedicated to gore and death contain perhaps the most alarming red flags. One comment claims that, as a child, the user watched her stepfather attempt to commit suicide. Another says she finds watching violent content addictive. A tracing tool shows that she had recently visited the profile of an American school shooter.

B.C. premier reacts

On Saturday morning, B.C. Premier David Eby released a statement, saying reports that Open AI “had related intelligence before the shooting” are “profoundly disturbing.”

“The pain that these families have gone through is unimaginable,” Eby said.

“We have confirmed with police that they are pursuing orders regarding the preservation of any potential evidence related to the shootings in Tumbler Ridge held by digital services companies, including social media platforms and AI companies.”

Eby encourages anyone with any information to let police know.

– With files from Jack Rabb and The Canadian Press