OpenAI tightens AI-safety rules after second account linked to Canadian shooter

OpenAI says it is overhauling safety protocols after a Canadian ChatGPT user who ran a second account allegedly used it in a mass shooting in British Columbia; the company had previously banned the shooter’s June 2025 account but did not report the case to police. In response, OpenAI plans a direct police contact in Canada, tougher measures to keep banned users from rejoining, and referrals to local resources for distressed or at-risk users. The changes follow talks with Canada’s government and BC officials and come as regulators push for stronger AI safeguards.