OpenAI tightens teen weapon rules and safety responses

OpenAI says it will introduce stronger limits on weapons-related information for teens, including practical details about weapons and security systems that could be misused.

The company also plans to publish its first Community Safety Report in early 2027, covering the second half of 2026, then issue updates every six months.

ChatGPT will further limit detailed reconstructions of past mass attacks and may end chats that show fixation on violence or seriously harming another person.