Abstract
This study explores the potential of generative AI to transform diary writing into a music-based creative activity, addressing the growing significance of AI-generated content in everyday cultural practices. While prior research has examined diary-based music composition and AI music interfaces, few studies have empirically evaluated the integration of generative AI music with diary content in terms of emotional response, service perception, and usability. To bridge this gap, this study aimed to design a conceptual AI-based diary app that converts personal diary entries into lyrics and automatically generates genre-specific music reflecting contextual data (e.g., date, weather, location, and emotion). An experimental evaluation was conducted with 73 participants who listened to four AI-generated songs composed from one month of diary entries and reviewed a prototype UI. Using a 5-point Likert scale, all key factors received positive evaluations above 4.0 (emotional enthusiasm: 4.39; emotional immersion: 4.31; sympathy: 4.21; convenience: 4.23; service immersion: 4.34; retention intention: 4.19). Multiple regression analysis revealed significant positive relationships between music immersion and service awareness (p < .01), with adjusted R² values ranging from 0.241 to 0.354. The System Usability Scale score averaged 86.13, exceeding the acceptability benchmark of 75. These findings indicate that diary-based AI music services can enhance emotional immersion and user retention through convenience and personalization. The study implies strong practical potential for generative AI-driven creative services and contributes empirical evidence for future AI-based cultural content design.