Abstract
BACKGROUND: Adolescent obesity remains a pressing public health challenge, particularly among socioeconomically disadvantaged populations. Artificial intelligence (AI) holds the promise for supporting students in managing daily health behaviors, but few existing studies used AI-based interventions in naturalistic settings such as schools. OBJECTIVE: This study evaluated the feasibility and preliminary impact of ProudMe Tech (Louisiana State University), an AI-assisted web app designed to help students manage 4 obesity-related behaviors: physical activity, screen time, diet, and sleep. METHODS: The 8-week, 1-arm pilot intervention study recruited 172 participants from 5 middle schools in Louisiana and used the ProudMe Tech to set behavior goals, track behaviors, record reflections, and receive AI-generated feedback. Both engagement (primary focus) and behavior impacts (secondary focus) were examined. RESULTS: Engagement metrics indicated varying levels of usage, averaging 8.9 (SD 7.6) behavior entries and 30.0 (SD 28.3) reflections per student, and receiving 33.5 (SD 29.7) AI feedback messages. Overall, participants recorded 6164 valid daily goals, of which 3934 (63.8%) were achieved. Natural language processing of the reflections and AI feedback messages revealed an overall neutral to positive sentiment. Pre- to postcomparisons showed (1) a significant reduction in screen time from 4.3 (SD 2.6) to 3.4 (SD 2.5) hours per day (21.6% decrease; t164=6.18, P<.001), (2) a small but significant decrease in fruit and vegetable intake from 5.7 (SD 3.8) to 5.2 (SD 3.5) servings per day (8.9% decrease; t169=2.27, P=.46), and (3) no significant changes in physical activity and sleep. CONCLUSIONS: These findings suggest that ProudMe Tech is a feasible AI chatbot that can engage adolescents in health behavior management, but more adaptation is needed to effectively elicit improvements in health behaviors and lower the obesity risk in middle school students.