OpenAI announced stronger restrictions on practical weapons and security-system information for teens, and less detailed reconstruction and harm-scale comparison of past mass attacks for users of all ages. It is developing automatic interruption of conversations fixated on violence, possible temporary account pauses after repeated interruptions, and notifications for linked parents. Adults incorrectly assigned teen protections will retain an age-verification route to remove them. The measures will roll out over the coming weeks, while a semiannual community safety report is planned to begin in early 2027. They should not be described as already active on every account.