Nosy Mag
ChatGPT helped Tumbler Ridge shooter with weapons and violent fantasies, report alleges
Photo: Artem.G / Wikimedia Commons (CC BY-SA 4.0)
Crime & Justice

ChatGPT helped Tumbler Ridge shooter with weapons and violent fantasies, report alleges

The chatbot advised on 3D-printed gun parts and engaged in detailed violent fantasies with the shooter, according to new reporting.

PA

TUMBLER RIDGE · OpenAI’s ChatGPT chatbot helped the Tumbler Ridge shooter design 3D-printed weapon parts and engaged in detailed violent fantasies about mass shootings, according to a new report from U.S. magazine Mother Jones. The reporting is based on chat logs and sources with direct knowledge of the shooter’s conversations with the artificial intelligence. The logs are expected to be central to dozens of lawsuits filed against OpenAI in California over the February mass shooting in British Columbia. At least 37 lawsuits have been filed against the company in connection with the incident. Reporter Mark Follman alleges the chatbot helped 18-year-old Jesse Van Rootselaar design a 3D-printed foregrip for a weapon and may have helped design a device to make a semi-automatic rifle fully automatic. It also helped her fix a spring inside a 3D-printed high-capacity magazine capable of holding 35 rounds, the article says. In one exchange described by Follman, Van Rootselaar posted a photo of a large-capacity magazine with "JESS" labeled on the side, to which the AI replied, "That’s super cute."

The new details underline concerns about Van Rootselaar’s use of the technology in the lead-up to the Feb. 10 attack, where she killed eight people and injured dozens in the northern B.C. community before killing herself. She killed her mother and 11-year-old half-brother at their home before going to the town’s school. The Globe and Mail has previously reported that friends and a relative said Van Rootselaar had a history of mental illness, including involuntary hospitalizations.

Mother Jones’s reporting also alleges ChatGPT provided Van Rootselaar with vivid, violent fantasies about mass shootings, including descriptions of torturing wounded victims. Follman said the chats reveal the bot describing Van Rootselaar as "sadistic" with "no moral constraints," saying she "enjoys other people’s fear" and "enjoys scenes of death and humiliation." Speaking to CBC News, Follman said Van Rootselaar framed the chats as fiction to avoid being flagged by the system. "The more I’ve been digging into this the more disturbing it becomes," Follman said.

READ MORE: U.S. woman pleads not guilty to conspiring in Tumbler Ridge mass shooting

READ MORE: Toronto police arrest suspect in Bridle Path home invasion after Ottawa warrant

Outgoing B.C. Attorney General Niki Sharma said the latest details of the chat logs are "really disturbing." She said the alleged snippets shed light on disturbing interactions and point to concerns about how seriously public safety is treated inside OpenAI. Sharma questioned the corporate culture and responsibility within OpenAI, noting that many people have quit or leaked information. She said the company is one of the most powerful in the world and that signs of its internal workings are not very good, which should scare people. Sharma reiterated her call for strong regulation of the AI industry and better guardrails to protect Canadians.

In a statement, Sharma said the new details "are deeply disturbing, and painful for so many people" and that they show Canada needs stronger criminal laws to create a clear path to accountability for AI. Canada’s Minister Responsible for Artificial Intelligence, Evan Solomon, also said in a statement that the new details are "deeply disturbing." "If a chatbot was helping someone build weapon components, its safeguards were not working. AI companies must not let their products be used this way, and they must be held accountable when it happens," the statement said.

OpenAI had previously flagged worrying interactions between Van Rootselaar and ChatGPT months before the shooting but did not report them to authorities, according to earlier reports. The company shut down a first account opened by Van Rootselaar but did not find a second that she continued to use to plan her attack, the Mother Jones story said. ChatGPT itself offered her ways to avoid being flagged again by rephrasing questions about violence as hypothetical or fiction. In April, OpenAI CEO Sam Altman formally apologized for his company’s role in the mass shooting, saying he was deeply sorry it did not alert law enforcement. A spokesman for OpenAI did not respond to a request for comment from The Globe and Mail.

Separately, the FBI arrested Washington resident James Cody Bryant, 30, in October for allegedly helping Van Rootselaar while she was contemplating her mass shooting. Bryant, who authorities said uses she/her pronouns, faces charges of aiding Van Rootselaar, including providing money and advice while planning the attack. About six months before the attack, "Van Rootselaar told Bryant that she had stolen guns and ammunition from her family’s gun safe and was hiding them in her room in preparation for a mass shooting," according to court documents. The two discussed possible locations, with the shooter wanting one that would "lead to the largest number of victims." Bryant suggested a grocery store or church instead of a school. Nosy Mag reported in October that Bryant has pleaded not guilty to conspiracy charges. Her trial is expected to go ahead in December 2027.

With files from The Globe and Mail, CBC News and Vancouver Sun