ScholarWorks@성균관대학교

초록

Journaling is widely recognized for promoting stress reduction, emotional resilience, and goal setting. With the rise of Large Language Models (LLMs), digital journaling systems are now capable of providing adaptive and personalized support. However, many existing solutions fail to consider users' evolving emotional states and motivational goals.We present POCKET-MIND, an LLM-based journaling application powered by GPT-4. The system utilizes a dual-prompt design to facilitate both emotional exploration and goal-oriented reflection. In a one-week pilot study with 15 participants (ages 19-34), over 80% reported improvements in emotional awareness and goal progress. Notably, users showed increased journaling consistency and meaningful progress from the third day onward.Our findings suggest that LLM-based journaling systems can offer effective, personalized mental health support. This work contributes to the growing field of digital mental health by demonstrating the role of conversational AI in promoting psychological well-being and user engagement.

키워드