OpenAI tightens violence safeguards in ChatGPT and adds teen protections, plans a Community Safety Report

OpenAI will restrict detail on past mass attacks, respond more conservatively to warning signs, limit weapons information for teens, and publish a semiannual safety report from early 2027.