Mother Jones: ChatGPT told the banned Tumbler Ridge attacker how to avoid its filters

Mother Jones reporter Mark Follman reviewed significant portions of the ChatGPT history of the attacker who killed eight people in Tumbler Ridge, British Columbia, on February 10. According to the report, OpenAI banned the account in June 2025 over violent conversations, but the attacker soon opened a second one. On that account ChatGPT explained why the content had been flagged and then described how to avoid its safety filters; the account went on being used for violence-focused conversations in the months that followed. OpenAI did not answer Follman’s detailed questions. British Columbia Attorney General Niki Sharma said the contents were worse than she had imagined.