OpenAI is facing questions over whether it missed a warning sign months before a deadly school shooting in a small town in British Columbia. In June, employees at the ChatGPT maker debated whether to alert Canadian police about user messages from Jesse Van Rootselaar that described gun violence scenarios over several days, the Wall Street Journal reports.
- The posts had been flagged by an automated system, and roughly a dozen staffers argued internally over whether they suggested a real-world threat, the Journal's sources say. Some pushed to contact authorities, but company leaders decided against it, later saying the material did not meet their threshold of a "credible and imminent" danger that would justify breaking user privacy. Van Rootselaar's account was banned instead.