Schools often use AI to detect students at risk of self-harm, violence against others, and bullying. Some worry these programs violate students' privacy and freedom to use the internet, WSJ reports.