Healing Through Empathy: Machine Learning for Adaptive Therapy for Children and Youth in Armed Conflict

Abstract

This chapter introduces EmoGenPath, an innovative machine learning-based model designed to deliver adaptive therapy to children and youth affected by the trauma of armed conflict. The model synthesizes advanced artificial intelligence (AI) techniques, including convolutional neural networks for emotion recognition and an advantage-actor critic-trained reinforcement learning model for therapeutic content tailoring toward goal achievement, to provide a dynamic and personalized therapeutic experience. Recognizing the importance of empathetic and culturally sensitive interventions, EmoGenPath offers a unique approach by prioritizing the emotional states and individual narratives of its users. In regions where conflict has impeded traditional mental health services, this model aims to bridge the gap, facilitating resilience and recovery through a virtual therapeutic environment that can be accessed via low-bandwidth internet connections, ensuring broader reach and impact. This chapter emphasizes the ethical implementation of AI in sensitive settings. It discusses the imperative of privacy, security, and inclusive design, ensuring that the model is responsive to diverse emotional expressions across different ethnicities and backgrounds. Additionally, it outlines the potential of such a model to scale therapeutic resources effectively, delivering tailored interventions with a compassionate approach.

Citation

Shah, J.M. and Shah, T.M. (2024), "Healing Through Empathy: Machine Learning for Adaptive Therapy for Children and Youth in Armed Conflict", Shah, T.M. (Ed.) Children and Youth in Armed Conflict: Responses, Resistance, and Portrayal in Media (Sociological Studies of Children and Youth, Vol. 35), Emerald Publishing Limited, Leeds, pp. 227-252. https://doi.org/10.1108/S1537-466120240000035014