OpenAI apologizes over police alert
- OpenAI chief Sam Altman apologized April 23 to Tumbler Ridge for not alerting police about a banned ChatGPT account later linked to Jesse Van Rootselaar. - OpenAI said it banned Van Rootselaar’s account in June 2025 after violent scenarios, but did not contact authorities before the February 10 school shooting. - Canada is weighing tougher AI safety rules after the killings and OpenAI’s response. (cbc.ca)
Sam Altman apologized to Tumbler Ridge after OpenAI failed to alert police about a banned ChatGPT account later linked to the February school shooting. (techcrunch.com) (apnews.com) In a letter dated April 23, Altman said he was “deeply sorry” that OpenAI did not tell law enforcement about the account it had banned in June 2025. (techcrunch.com) (usnews.com) Police identified 18-year-old Jesse Van Rootselaar as the attacker in the February 10 killings in Tumbler Ridge, British Columbia. CBC reported that eight victims were killed before Van Rootselaar died by suicide. (cbc.ca) (apnews.com) OpenAI said its systems and human reviewers had flagged the account for possible violent misuse, then suspended it. The company said the material did not meet its internal threshold for an “imminent and credible risk” requiring a law-enforcement referral. (cbc.ca) The case turned a technical policy question into a public one: when a chatbot company sees violent planning, who gets called and how fast. CBC said federal and provincial officials argued the company should have escalated sooner. (cbc.ca 1) (cbc.ca 2) British Columbia Premier David Eby said Altman’s apology was “necessary” but “grossly insufficient” for families in Tumbler Ridge. Altman said he had spoken with Eby and Mayor Darryl Krakowka before issuing the letter. (techcrunch.com) OpenAI says it is changing its procedures by using more flexible criteria for referrals to authorities and setting up direct points of contact with Canadian law enforcement. (techcrunch.com) (cbc.ca) Canada’s artificial intelligence minister, Evan Solomon, said in February that he was “deeply disturbed” by the delayed reporting and that Ottawa was reviewing “a suite of measures” to protect Canadians, especially children. (cbc.ca) No new Canadian rules have been announced yet, but the apology has kept the focus on whether AI companies should treat some chatbot conversations like emergency warnings. OpenAI’s answer, for now, is that its old threshold was too narrow. (techcrunch.com) (cbc.ca)
