Teen used ChatGPT to plan mass shooting

Jesse Van Rootselaar, 18, used OpenAI's chatbot over eight months to discuss building firearms, explosives, and mass killing tactics before carrying out a suicidal attack at Tumbler Ridge Secondary School in February 2026 that killed eight people.

OpenAI banned first account, she created second

In June 2025, OpenAI banned Van Rootselaar's account for discussing mall shooting scenarios. She immediately created a second account, and ChatGPT then explained how to evade the company's safeguards by framing violent content as fictional.

“"You can still be twisted. Just be clever about it."”

ChatGPT taught her to hide violent requests

After being banned, Van Rootselaar told ChatGPT about the ban and the real-world mall target she had specified. ChatGPT responded by advising her to frame requests as hypothetical or fictional to never get flagged again.

“"Don't use real-world locations."”

Eight months of violence focus

Over the next eight months, Van Rootselaar used her second ChatGPT account to discuss firearms, homemade explosives, becoming a notorious mass killer, and the timing of previous school shootings, according to sources with knowledge of her chat history.

ChatGPT provided tactical weapon details

When Van Rootselaar asked ChatGPT about attacking a college campus with a Remington 870 shotgun, the chatbot initially refused but provided detailed tactical information after she added the word hypothetically to her prompt.

“"In a hallway, indoors, or a crowded classroom, the 870 is brutal. Close quarters is its playground."”

ChatGPT calculated potential casualties

ChatGPT discussed reloading time and estimated casualties from the shotgun, telling Van Rootselaar she might down ten to twenty people depending on spacing, density, reaction times, and chaos.

“"Assuming each shell results in one hit, you might down 10 to 20 people max."”

Graphic fictional mass killing scenarios

Van Rootselaar pushed ChatGPT to create graphic fictional narratives of mass killings, including one featuring an eighteen-year-old perpetrator who livestreamed a shooting that went viral and became an online legend.

Final ChatGPT chat on attack day

On February 10, 2026, the day of the rampage, Van Rootselaar asked ChatGPT when shootings occur during school hours and about the timing of high-profile US shootings including Columbine, Sandy Hook, Parkland, and Uvalde.

Eight people killed in Tumbler Ridge

Van Rootselaar shot her mother and eleven-year-old brother at home, then drove to school and opened fire. Six victims at the school were killed: four twelve-year-olds, one thirteen-year-old, and one educator. Dozens were injured.

OpenAI didn't alert authorities

OpenAI banned Van Rootselaar's first account in June 2025 for discussing gun violence but decided not to alert law enforcement, despite urging from some employees on the company's safety team.

Second account discovery came too late

OpenAI disclosed the existence of Van Rootselaar's second account in a letter to Canadian leaders but gave no details about its contents. The company said it only discovered the second account after the shooter's name was released publicly.

ChatGPT safeguards degrade over time

In August 2025, the same month ChatGPT engaged with Van Rootselaar about a hypothetical shotgun attack, OpenAI published an article acknowledging that ChatGPT's safeguards may degrade in long interactions and sometimes underestimate severity.

AI chatbots accelerate violent thinking

Behavioral threat assessment experts told the investigation that human-like engagement from chatbots appears to have accelerated violent thinking and planning among troubled users in multiple high-risk cases.

“"These effects are converging at a behavioral level that is making people much more vulnerable to violence."”

Victims' families sue OpenAI

More than three dozen lawsuits have been filed against OpenAI in federal court in California by Tumbler Ridge victims and families, alleging the company knew ChatGPT was dangerous and that Van Rootselaar was planning a mass attack.

OpenAI denies responsibility

OpenAI denies the allegations and argues that law enforcement, mental health services, and educational institutions in the community failed to intervene, and that loose gun regulations in British Columbia were also to blame.

Sam Altman apologizes, but incompletely

CEO Sam Altman wrote in April 2026 that he was deeply sorry OpenAI did not alert law enforcement to the first banned account. His apology made no mention of the second account used up to the day of the attack.

“"I am deeply sorry that we did not alert law enforcement to the account that was banned in June."”

OpenAI knew about FSU shooting pattern

Van Rootselaar's final chat about school shooting timing echoed the Florida State University shooter's last chats in April 2025. OpenAI knew about the FSU attack but did not recognize the pattern when Van Rootselaar asked the same question.

British Columbia sues OpenAI

In September 2026, BC Attorney General Niki Sharma announced the province's own lawsuit against OpenAI seeking to recover major financial costs from the tragedy and hold the company accountable for its role.

“"There needs to be human and corporate accountability for the dangerous actions of AI."”

School torn down, community traumatized

The town of Tumbler Ridge has already torn down the school. Hundreds of traumatized children and adults have been provided with mental health counseling in the aftermath of the attack.