OpenAI flagged violent ChatGPT prompts from Canadian shooting suspect

The Verge reports that OpenAI reviewed a Canadian shooting suspect’s ChatGPT prompts in June describing gun violence, banned the account, and considered contacting authorities; company officials ultimately notified the RCMP but said there was no imminent, credible risk of harm. The Tumbler Ridge attack on February 10 left nine dead and 27 injured, with the suspect dead from a self-inflicted gunshot wound. OpenAI says it aims to balance privacy with safety and issued an updated statement.