OpenAI is enhancing its safety measures after an internal flagging of a suspect's ChatGPT account failed to prevent the deadly Tumbler Ridge shooting. The company acknowledged that the suspect, Jesse Van Rootselaar, created a second account after his initial one was banned in June 2025, bypassing internal safeguards.
Canadian officials criticized OpenAI for not reporting the account to police, suggesting the attack, which claimed eight lives including the suspect's mother and stepbrother, could have been averted. OpenAI stated that under its revised guidelines, the suspect's activity would now be reported to law enforcement, citing improved criteria and collaboration with mental health experts.
Canada's AI Minister Evan Solomon expressed disappointment, noting a lack of substantial new safety protocols and leaving open the possibility of future legislation. British Columbia Premier David Eby believes timely reporting could have prevented the tragedy and confirmed OpenAI CEO Sam Altman will discuss safety policies.