Schools considering AI-powered mental health tools face rising scrutiny after a new risk assessment warned that the market is largely unregulated and some apps can be harmful to teens. Common Sense Media, working with Stanford University’s Brainstorm Lab, assessed both “institutional” school-based AI mental health apps and direct-to-consumer products. The assessment found school-based apps performed better overall on risk and helpfulness checks, including how they detect warning signs, estimate severity, provide crisis resources, and avoid advice that could worsen symptoms. Researchers also reported that three in 10 teens have used an AI mental health app, and even more have used general-purpose AI like ChatGPT for emotional support. The assessment’s key caution: these apps can complement human services, but cannot replace the staffing and clinical support students still need.