LLM-Based Interview Bot for Student Big Five Assessment and Career Recommendation

Abstract

The development of Artificial Intelligence (AI) and Natural Language Processing (NLP) offers new opportunities to make psychological assessments more interactive and meaningful. However, personality tests such as the International Personality Item Pool – Big Five Factor Markers (IPIP-BFM-50) still rely on static self-report questionnaires, which may limit engagement and contextual interpretation. This study proposes an InterviewBot-based Big Five Personality system (IB-B5P) that combines rule-based IPIP scoring with Large Language Model (LLM)-driven conversational assessment using GPT-3.5 Turbo. The system generates both quantitative personality scores and qualitative narrative profiles. Evaluation results show moderate to strong correlations (r = 0.31–0.71) between IB-B5P and IPIP scores, with Openness and Extraversion showing statistically significant relationships. These findings suggest that the hybrid rule–LLM approach can approximate IPIP tendencies while providing richer context-aware interpretations. The novelty of this study lies in integrating LLM-based conversational intelligence with a standardized psychometric framework, with potential applications in career guidance, educational counseling, and digital psychological assessment in higher education.