Schools are increasingly considering AI-powered mental health apps amid psychologist and counselor shortages, but a new risk assessment warns the market is largely unregulated and some tools can be harmful. Common Sense Media, alongside Stanford University’s Brainstorm Lab, evaluated institutional (school-based, human-in-the-loop) and consumer AI mental health apps. The assessment found that institutional apps that keep humans involved performed better, while direct-to-consumer products raised greater concerns about whether apps can recognize warning signs, assess severity accurately, provide crisis resources, and avoid advice that delays care. Separate research cited in the report indicates about three in 10 teens have used an AI mental health app, and more have used general-purpose AI for emotional support. The message for higher education and teacher-prep programs is clear: as K–12 districts experiment, educators need procurement, monitoring, and safeguarding frameworks for student-facing AI tools.