<p>As artificial intelligence (AI) becomes deeply embedded in higher education, understanding its psychological and cognitive impact on the writing process is critical. However, limited research has examined the psychological and learning mechanisms through which AI influences students’ writing development. This study aims to examine the association between AI integration and student writing outcomes. A quantitative, cross-sectional design was employed; data were collected from 340 university students via a convenience and snowballing sampling technique. Exploratory analysis identified seven factors: writing motivation and engagement, writing independence and autonomy, creative expansion, perceived risks and challenges, trust and perceived objectivity, functional workflow integration, and critical and structural reflection. Confirmatory factor analysis further demonstrated satisfactory reliability and construct validity. Structural equation modeling revealed that functional workflow integration, trust and perceived objectivity, and perceived risks and challenges significantly influenced writing outcomes through the mediating roles of creative expansion and writing independence and autonomy. While trust and perceived objectivity directly enhanced writing motivation, functional workflow integration primarily affected motivation indirectly through creativity and autonomy. Creative expansion emerged as a key pathway linking AI use to students’ critical and structural reflection. The model explained 54% of the variance in writing motivation and 36% in critical reflection. These findings suggest that AI acts as a cognitive scaffold that enhances student writing only when it satisfies the learner's need for autonomy and creative broadening. The study concludes that pedagogical strategies should shift from monitoring AI usage to facilitating its role as a mediational tool for self-regulated writing.</p>