Does AI predict personality traits? A meta-analytic review of machine-predicted Big Five scores: Convergence and caveats

The validity of AI prediction on personality remains underexplored. In this study, we examine the convergence between machine-predicted and self-reported Big Five scores through two integrated multilevel meta-analyses(K=121) addressing both regression and classification tasks. Our results reveal convergent validity, with pooled mean correlations ranging from .38 to .44 (95% CI: .16–.58) and mean accuracy scores between 65% and 70% (95% CI: 60%–74%). The length of the personality questionnaire emerged as a significant moderator of between-study heterogeneity. We compared the factor loading structures between machine-predicted and self-reported scores using meta-SEM. The high factor congruence (.9) indicates that machine-predicted scores may be exposed to common method bias akin to self-reports. These findings underscore cautious applications of AI-driven personality assessments in real-world settings.