Automating personality-based employment interviews: A systematic review, development and validation of an AI chatbot, and practical considerations

Abstract

Advancements in Artificial Intelligence (AI) and natural language processing (NLP) have enabled the development of innovative methods to assess personality using textual responses elicited from structured behavior-based interviews, chatbot-based open-ended interviews, and automated video interviews. This dissertation aims to advance the field by conceptually integrating literature on automated, personality, and employment interviews and developing an AI chatbot that administers and assesses personality through situational personality-based employment interviews, providing practical utility for selection purposes. This dissertation comprises three main studies. In Part I, I conducted a systematic literature review of research on employment, personality, and automated interviews to synthesize existing findings, identify key themes and gaps, and explore methodological advancements. Results from structural topic modeling uncovered five latent themes and highlighted overlap and underrepresented themes areas within the domains of internet research. In Part II, I developed and validated an AI chatbot to administer a personality-based employment interview, leveraging word embeddings and zero-shot prompt engineering to assess personality from participants’ textual responses. Results demonstrated moderate convergence between chatbot-derived and self-reported personality scores, though response quality (e.g., brevity in chatbot interviews) influenced scoring accuracy. In Part III, I examined the practical considerations of implementing AI-based selection tools, including user perceptions, subgroup differences, and resistance to response distortion. Results revealed mixed support for chatbot-based assessments in selection contexts. While participants found the chatbot engaging, concerns remained regarding usability, fairness, and faking resistance.