Designing for mental wellbeing: exploring the role of artificial intelligence for reflective practice

Abstract

The advancements in personal data collection through sensor and artificial intelligence (AI) technology have opened new opportunities for individuals to engage in self-reflection and introspection. In the field of mental health, these technologies have augmented traditional, subjective, and intuition-based evaluations of personal emotions with more objective, data-driven approaches. However, the integration of AI may potentially reduce the active role of the user in interpreting and finding meaning in the collected data. This study aims to investigate the potential for AI capabilities and roles, such as prediction, conversation, and generation, to be used in a collaborative manner with the user during the process of self-reflection. To this end, I conducted the following studies: (1) the design of a personal informatics system utilizing a stress prediction model and explainability to collaborate with the user in understanding, reflecting on, and intervening in stress levels and causes, (2) the design of interactions between human-conversational agents to assist users in re-examining their cognitive biases and related past events, and (3) the exploration of the potential of large-scale language generation models to facilitate self-reflection and introspection through collaborative diary writing with the user. Finally, this research proposes a design opportunity for using AI in the process of self-reflection to improve mental health.