OpenAI limits weapons information for teens on ChatGPT and will let it end chats that fixate on violence, Reuters reports

OpenAI is tightening the weapons-related information ChatGPT gives users under 18, Reuters reported on 30 October. The limits cover practical information about weapons and security systems even when a request shows no sign of violent intent.

For all users, ChatGPT will be able to end conversations in which someone appears to be fixating on violence, and OpenAI may pause an account temporarily after several such chats, Reuters reported; Malay Mail also reported the chat-ending systems and possible suspensions. Reuters said OpenAI plans to notify parents if their teen's account is paused for repeatedly discussing violence.

Malay Mail reported that OpenAI is also limiting detail about past mass attacks, including the weapons used and how they were obtained, and that adults wrongly classed as minors by its age prediction can verify their age. Both outlets said the changes roll out over the coming weeks.

OpenAI's announcement is titled “Community safety report and new protections for teens”. Its page refused us twice, so this story rests on the two outlets. It follows Texas's lawsuit over ChatGPT's safety, which the Switchboard ran on 25 October, and the 27 October suit by parents over medication advice GPT-4o gave their daughter (in Catching up below).

• Reported OpenAI will limit weapons information for under-18s, let ChatGPT end violence-fixated chats, may pause accounts, and plans to notify parents when a teen's account is paused, Reuters reported. Reuters, via Investing.com
• Reported OpenAI is also limiting detail about past mass attacks, and adults misclassified as minors can verify their age, Malay Mail reported. Malay Mail, via Yahoo News Malaysia