Meta has now built a system that scans teens' chats with Meta AI for signs of self-harm. It was developed with the help of parents and mental health experts to identify conversations where a teen may be at risk, even if the language is subtle. Stricter content settings will further limit inappropriate conversations for young users.