Some AI mental health apps are harmful for kids, says report—what experts say parents should keep in mind A report by Common Sense Media evaluated AI mental health apps for teens, finding that school-deployed apps like Alongside and Sonar posed lower risks than direct-to-consumer apps like Wysa. School-based apps are safer due to integrated human oversight and connections to existing support systems, which can escalate concerns appropriately. Direct-to-consumer apps often miss critical signs and provide inadequate support, presenting an unacceptable risk level for teen users.

• General purpose chatbots like ChatGPT and Claude are used by teens for social practice and emotional support.
• Common Sense Media’s report found direct-to-consumer AI mental health apps have an ‘unacceptable’ risk level for teens.
• School-deployed AI mental health apps Alongside and Sonar showed ‘low’ or ‘minimal’ risk levels.
• Direct-to-consumer apps missed signs of serious mental health conditions and failed to refer users to crisis lines.
• School-based apps are safer because they involve real people and connect to school support systems, flagging concerning activity.
• Even school-based apps sometimes missed signs for issues like eating disorders.
• Direct-to-consumer apps lack structural features that make school-based apps safer.
• Experts advise parents to have open conversations with their children about app usage and to inquire about school systems’ mental health support.