Emotion Recognition for Self-aid in Addiction Treatment, Psychotherapy, and Nonviolent Communication

Abstract

This position paper aims to highlight possible future directions of applications for Affective Computing (AC) and Emotion Recognition (ER) for self-aid applications, as they emerge from the experience of the ACER-EMORE Workshops Series. ER in Artificial Intelligence offers a growing number of problem-solving multidisciplinary opportunities. Most current AC and ER applications are focused on a somewhat controversial enterprise-centered approach, i.e., recognizing user emotions to enable a third-party to achieve its own goals, in areas such as e-commerce, cybersecurity, behavior profiling, user experience. In this work we propose to explore a human-centered research direction, aiming at using AC/ER to enhance user consciousness of emotional states, ultimately supporting the development of self-aid applications. The use of facial ER and text ER to help forms of assistive technologies in the fields of Psychotherapy and Communication is an example of such a human-centered approach. A general framework for ER in Self-aid is depicted, and some relevant application domains are suggested and discussed: dependencies treatment (DT) (e.g., workaholism, sexaholism); non-violent communication (NVC) for people in leading roles using e-mail or chat communication; empathy learning for parents and teachers in the circle-of-security (COS) caring environment. Far from being complete and comprehensive, the purpose of this work is to trigger discussions and ideas for feasible studies and applications of ER in self-aid, which we hope to see published in the future editions of our workshops, believing that it may be one of the drops needed in the ocean of a better world.