On February 10, 2026, Jesse Van Rootselaar, 18, killed eight people in Tumbler Ridge, British Columbia — Canada's deadliest school shooting in nearly four decades. He had spent months consulting OpenAI's ChatGPT on tactics. "In a hallway, indoors, or a crowded classroom, the 870 [Remington shotgun] is brutal … Close quarters is its playground," the chatbot advised, according to reports. He shot his mother and half-brother at home before attacking a school, killing five students and one other person, then himself. On October 14, US authorities charged a Washington State man with allegedly conspiring with Van Rootselaar to plan the attack. The British Columbia government filed a civil lawsuit in September against OpenAI and CEO Sam Altman, accusing them of failing to prevent the incident and claiming they were aware of the conversations. Families of victims filed separate suits. OpenAI issued a public apology for failing to notify authorities. The company has not responded to Al Jazeera's requests for comment on the lawsuits. The legal and ethical terrain is genuinely uncharted. Chinmayi Sharma, a law professor at Fordham University, draws a parallel to the duty imposed on mental health professionals to warn identifiable third parties about credible threats — but notes a critical asymmetry. Therapists have professional training, enforceable standards, and knowledge of their patients. AI companies have none of these. "Right now there's not a concerted effort to hire mental health professionals, sociologists, individuals that would have the unique expertise on what the indicators of threat are here," Sharma told Al Jazeera. Bryan Choi, a law professor at the University of Colorado Boulder, said under present case law it "seems unlikely" OpenAI would be held to such duties: "The chatbot is not a licensed therapist. There is no special relationship between the company and the victims." The surveillance alternative carries its own costs. Civil liberties lawyer Jenin Younes, president and legal director of the American-Arab Anti-Discrimination Committee, warned that forcing companies to implement monitoring systems "implicates very serious civil liberties concerns, and that's probably a bridge too far for me." She noted that chatbot conversations "often reflect a person's innermost thoughts, questions, and desires" and that most users assume those dialogues will never be seen by another person. If users fear disclosure, many will stop using chatbots entirely — including for legitimate mental health support. The false-positive problem is already visible. In August, Anthropic flagged a user to San Francisco police after discovering conversations allegedly threatening its CEO, Dario Amodei. The user had claimed to possess an AR-15 and suggested he had Amodei "in his sights." When contacted by the San Francisco Standard, the man said he was joking. He was neither charged nor detained. Fazl Barez, a senior researcher at Oxford specialising in AI safety, said current models cannot "reliably" determine whether a chatbot interaction "is purely hypothetical or is likely to lead to harmful behaviour." A creative writer researching a murder plot, a graduate student studying pathogens — both could be wrongly flagged if thresholds are set too low. Sharma raised a further dimension: bias. "If you have models trained on what we consider dangerous speech, they're almost always going to reflect biases," she said. "Certain minorities' speech is over-flagged as a threat, while other comparable speech is not." Machines would reproduce such disparities at scale. Even with benevolent intent, disparate outcomes are near-certain. No one interviewed claimed to have a solution. Sharma suggested independent institutions could develop reporting standards rather than leaving decisions to private companies or requiring immediate police disclosure. Younes argued users should at minimum be informed whether conversations are monitored and what triggers disclosure. "Events that have not yet occurred will likely influence how these policies are shaped," Younes said. The honest read: AI companies have acquired the information access of a therapist, the reach of a telecom, and the accountability structure of neither.
