Big Five Personality Traits in AI-Generated Avatars: Digital Self Representation Among Malaysian Gen-Z

Abstract

Generative artificial intelligence has transformed how individuals construct digital self-representations, especially with personalised AI-generated avatars. However, little is known on how these avatars align with users’ personality traits in non-Western contexts, in particular among Malaysian Gen-Z undergraduates as they are highly immersed in social media and virtual environments. This study examines the relationship between Big Five personality traits and the selection, perceived accuracy and attitudes toward AI-generated avatars for this demographic. A mixed-methods design was employed: 97 undergraduates completed a validated Big Five personality assessment and, using a custom-built online platform (“Gensona”), selected from a set of AI-generated avatars tailored to their dominant traits, followed by a post-selection survey. Semi-structured interviews with a small group of participants provided deeper insight into the meanings they attached to their avatar choices and evaluations. Quantitative findings indicate that Agreeableness and Extraversion are most strongly associated with personality-driven avatar selection and more positive attitudes toward AI-generated avatars, while overall acceptance and perceived accuracy are moderate across the sample. The qualitative findings also show that students use avatars to represent who they perceived they are, and also to explore ideal or context-specific versions of themselves. This reveals tension between authentic, personal aesthetic choices and expectations set by different digital platforms. Furthermore, the study highlights the potential and limitations of AI-generated avatars in representing genuine digital self. It also points out the importance of design and ethical considerations for future AI-driven visual identity systems aimed at Malaysian Gen-Z users.