Enhancing Mental Health Support for Young Adults Through Conversational AI-Based Screening Tools

Abstract

Young adults are increasingly vulnerable to depression and anxiety due to educational, social and
economic pressures, but stigma and limited access to care often prevent early intervention. This
study evaluates the acceptability and performance of a conversational AI-based screening tool that
integrates the validated Patient Health Questionnaire-9 (PHQ-9) and Generalized Anxiety
Disorder-7 (GAD-7) scales. Using a cross-sectional experimental design, data were collected from
young adults aged 18–35 interacted with the natural language chatbot. A total of 950 young adults
participated in this study, contributing conversational text and standardized scores. Preprocessed
responses were analyzed using a support vector machine (SVM) classifier to distinguish between
crisis (+1) and non-crisis (-1) expressions. The model achieved a precision of 84.07%, recall of
83.62%, and F1 score of 83.84%, confirming its reliability in identifying emotional distress.
Participants reported greater comfort and openness when communicating with chatbots compared
to traditional self-report formats, indicating increased engagement and reduced stigma. The
findings highlight the potential of conversational AI to improve early mental health screening and
provide accessible, non-judgmental support. The study recommends ethical integration of AIpowered
screeners into consultation and telemedicine systems to expand scalable, privacypreserving
mental health care for young adults.