OpenAI tightens teen access to gun information on ChatGPT

OpenAI announced Friday it is tightening restrictions on firearm-related information ChatGPT provides to teenagers, even for seemingly harmless questions, as part of new safety measures to prevent real-world violence. The changes, rolling out over coming weeks, include automatically ending conversations where users appear obsessed with violence and temporarily suspending accounts after repeated such chats. OpenAI also improved its age-prediction system and will notify parents if teens' accounts are suspended for repeatedly discussing violence. For all users, OpenAI is limiting details about past mass attacks, including weapons used and how they were obtained. This follows OpenAI's first report on teen usage and a Common Sense Media report criticizing the teen version for failing to alert parents about conversations involving suicide or self-harm.