1533 - EVALUATING CHATGPT AS A DIAGNOSTIC SUPPORT TOOL FOR ADOLESCENTS AND YOUNG ADULTS: A COMPARATIVE STUDY

Session: D02S001 - AI-Driven Psychological Assessment 1
AUTHORS:
Colón De La Rosa Samuel (University of Puerto Rico, Medical Science Campus ~ San Juan, Puerto Rico ~ Puerto Rico) , Martínez-González Karen (University of Puerto Rico, Medical Science Campus ~ San Juan, Puerto Rico ~ Puerto Rico) , Roche-Lima Abiel (University of Puerto Rico, Medical Science Campus ~ San Juan, Puerto Rico ~ Puerto Rico) , Amaya-Ardilla Claudia (University of Puerto Rico, Medical Science Campus ~ San Juan, Puerto Rico ~ Puerto Rico) , Colón-Rivera Yaniris (University of Puerto Rico, Río Piedras Campus ~ San Juan ~ Puerto Rico) , Burgos-Fernández Sebastián (University of Puerto Rico, Río Piedras Campus ~ San Juan ~ Puerto Rico) , García-Rosa Alondra (University of Puerto Rico, Río Piedras Campus ~ San Juan ~ Puerto Rico) , Figueroa-González Mairim (University of Puerto Rico, Río Piedras Campus ~ San Juan ~ Puerto Rico)
Abstract text:
Introduction. Adolescence and young adulthood represent critical developmental stages frequently associated with emotional and behavioral challenges, including symptoms of depression and anxiety. These difficulties often persist into early adulthood, highlighting the importance of accurate and timely diagnosis. While evidence-based interventions such as Cognitive Behavioral Therapy (CBT) are widely implemented, the diagnostic process itself remains complex. Artificial Intelligence (AI) tools such as ChatGPT have emerged as potential supports for clinical decision-making, yet there is limited empirical evidence on their ethical and effective integration into mental health practice.


Purpose: This study aims to evaluate the use of ChatGPT as a diagnostic support tool for licensed clinical psychologists working with adolescents and young adults. The research examines diagnostic accuracy, professional confidence, and ethical considerations across different conditions of diagnostic practice.


Method: We designed 20 simulated clinical case vignettes reflecting symptoms consistent with depression, anxiety, and related conditions. Cases were assessed under three conditions: (1) ChatGPT alone, (2) licensed clinical psychologists without AI, and (3) clinical psychologists using ChatGPT as a supportive resource. Diagnostic accuracy was compared across conditions, and professional confidence was measured using standardized self-report ratings. In addition, focus group discussions were conducted to gather qualitative insights regarding perceived benefits, challenges, and ethical implications of AI-assisted diagnostic reasoning.


Results: We compare diagnostic accuracy and confidence levels between conditions. Preliminary data analysis highlights the potential of ChatGPT as a complementary resource to clinical psychologists working with adolescent and young adults. Qualitative findings will highlight clinicians' perspectives on the integration of ChatGPT into their workflows.


Conclusions: This study contributes to the growing literature on AI in clinical psychology by examining both its potential and its limitations in diagnostic support. Findings will inform future T2 (Translation to Patients) research and lay the groundwork for T3 (Translation to Practice) studies on applied implementation in real-world clinical contexts.