ABSTRACT
Background: AI has brought new opportunities for mental health treatment; among them is the application of chatbots based on Artificial Intelligence. They also represent the conversational agents that are programmed to act like human companions to provide immediate intervention encouragement and even evaluate a user’s mental health status. Its use in mental health care is the most significant diagnostic tool in increasing accessibility, early detection/ identification, and individualized treatment of mental health diseases. The study examined the diagnostic potential of AI-powered chatbots in mental health care, assessing their ability to detect symptoms of mental health disorders and provide appropriate interventions or referrals.
Methods: The study focuses on Breezy, an AI-powered chatbot developed to engage users in social conversations while addressing depressive symptoms, set to be operational by 2025. It is targeted at young Nigerians, Breezy will converse in both English and Pidgin, providing a unique mental health support service. Breezy’s capabilities include understanding and responding to user inputs, with a feature that explains how users can answer specific questions. Breezy stands out as Nigeria’s first and only open-domain mental health chatbot, meaning it can discuss a wide range of topics beyond depression. This versatility enhances its realism and effectiveness as a tool for understanding how individuals seek social support. Breezy will be accessible via a website and mobile app, ensuring broad reach and engagement. The chatbot’s design was inspired by datasets from Kaggle, which include]ed sentences containing depression-related keywords like
“depressed,” “sad,” “feeling low,” and “hate me.” Breezy is not designed for clinical diagnoses but aims to provide social support to users experiencing depressive feelings. It will be available as a premium service across Nigeria, offering a new way to address mental health issues through accessible AI technology.
Results: The primary distribution of the intent of the chatbot was “sleep”, “about”, “sad”, “depressed”, “learn more”, “help”, “something else”, “scared”, “user-meditation”, “thanks”, These keywords are depressive (containing depression- or sadness-related keywords), happy (containing happiness- or excitement-related keywords), and general (indicating all utterances in the same period).
Conclusion: This study adds to the discussion on the application of AI chatbots in mental health care. It provides valuable knowledge about how the usage of such technology can transform the diagnostic approach and try to improve global access to mental health services. The findings show that chatbots may share valuable information about depressive states, particularly in users who cannot describe their feelings to others.
How to cite: Odunuga, K., Akinyemi, K., Adeniran, KO. Igboke, HC. AI-Powered Conversations: The
Diagnostic Potential of Chatbots in Mental Health Care. Global Health Professionals Multidisciplinary
Practices Journal November 2024, 1(4):25-33
DOWNLOAD THIS DOCUMENT
