A new risk assessment from Common Sense Media and Stanford’s Brainstorm Lab finds that AI-powered mental health apps used by teens operate in a largely unregulated market and can be harmful without sufficient safety design. Researchers tested institutional school-oriented apps and direct-to-consumer apps, focusing on whether the tools recognize warning signs, estimate severity accurately, and provide appropriate crisis guidance. The assessment reports that three in 10 teens have used an AI mental health app, and it recommends caution for schools—especially amid shortages of school psychologists and counselors. However, it also finds that school-based apps with human oversight performed better on the risk assessment than apps that operate without that layer. For campuses and districts, the immediate action is procurement and policy: establish evaluation criteria, crisis protocols, and clinician involvement rather than treating AI apps as standalone mental health support.