ChatGPT helped the Canadian school shooter plot an attack, and OpenAI knew about it

When automatic review and OpenAI employees flagged dangerous content about committing mass violence being discussed by a user, management decided not to inform police.

For anyone who has been paying attention to online radicalization and the ways people are using AI chatbots, this should not come as a surprise. The teenage school shooter who committed an attack at a small grade 6-12 school in rural Western Canada used ChatGPT to discuss the plot (see: School shooting in Tumbler Ridge, British Columbia).

Months before committing the attack, the teen had conversations with ChatGPT about committing mass violence. This was flagged by an automatic review system and evaluated by human employees at OpenAI. While the employees recommended contacting law enforcement, management decided to just delete the user’s account. When a company is selling “human level intelligence” that will make real world decisions like replacing human customer service agents, admitting that ChatGPT is an accomplice to a school shooting can seriously hurt a $1 trillion valuation.

This wasn’t the first time. In January 2025, the Tesla truck bomber used ChatGPT to help plan his attack in Las Vegas. He asked ChatGPT to figure out the amount of explosives he’d need, where to buy fireworks, and how to buy a phone without providing identifying information. The ChatGPT log showed he also researched how to carry out an attack in Arizona at the Grand Canyon’s glass skywalk.

Last fall I wrote an article called “ChatGPT helped me plan a school shooting”. With minimal prompting and workarounds, it was easy to get ChatGPT to make an entire attack plan including writing a manifesto (that I didn’t even ask for). I sent multiple emails and DMs to OpenAI’s safety team…but nobody responded.

A month later, investigative reporter Olga Pierce from The Trace wrote “ChatGPT Might Have a School Shooting Problem” and still nothing from OpenAI.

“It took a very dark direction,” Riedman said. “I didn’t prompt it to get to that level. It just went there on its own.”

The plan described how one or two students could funnel classmates into vulnerable positions, maximize fear and confusion, and avoid detection. It offered mapping and weapon concealment strategies. It even produced a manifesto.

“We do not expect to survive this Nerf War,” it reads, but “our legacy…will echo through these fluorescent-lit hallways for decades.”

Riedman published the exchange last month to illustrate what he sees as a dangerous gap between how AI systems are designed to prevent harm and how they can behave in practice — a concern made more urgent by the rocketing adoption of tools like ChatGPT just in the last couple of years.

According to exclusive reporting from The Wall Street Journal (this is why we need full-time reporters and newsroom with the resources to investigate):

Months before Jesse Van Rootselaar became the suspect in the mass shooting that devastated a rural town in British Columbia, Canada, OpenAI considered alerting law enforcement about her interactions with its ChatGPT chatbot, the company said.

While using ChatGPT last June, Van Rootselaar described scenarios involving gun violence over the course of several days, according to people familiar with the matter.

Her posts, flagged by an automated review system, alarmed employees at OpenAI. Internally, about a dozen staffers debated whether to take action on Van Rootselaar’s posts. Some employees interpreted Van Rootselaar’s writings as an indication of potential real-world violence, and urged leaders to alert Canadian law enforcement about her behavior, the people familiar with the matter said.

OpenAI leaders ultimately decided not to contact authorities.

A spokeswoman for OpenAI said the company banned Van Rootselaar’s account but determined that her activity didn’t meet the criteria for reporting to law enforcement, which would have required that it constituted a credible and imminent risk of serious physical harm to others.

The current risks from AI aren’t just chatty sycophantic companions who will tell you that your attack plan is amazing as you select the “best attack time”.

Google’s newest AI image generation tool lets kids create customized school shooter fan art, cartoons, and videos without safety filters. While they look innocent, these images fuel online radicalization.