AI面試系統及個人評估問卷對於人格特質評價之練習效應

隨著互動式資訊技術的迅速發展，非同步視訊面試因其快速且便利的特性，已更加被企業所接受並使用，透過結合人工智慧及機器學習的技術，即可分析個人的五大人格特質。同時，也有研究證實人格特質能夠預測未來職場表現，但應徵者是否能過不斷的練習，來改變過去常被企業使用的自我評估問卷或是人工智慧面試系統所分析出的人格特質？若是求職者可以透過練習改變測評結果，表示求職者能夠扭曲該測評方式的效度，該測評方式也不具信度。因此，本研究將探討以自我評估問卷及人工智慧面試兩種方式在評估人格特質，受試者所展現的練習效應。評估人格特質自我評估問卷與人工智慧評估人格特質之間，何者具有更高的可靠性。本研究採用準實驗研究法，研究對象為個案公司中的120位企業內部員工。結果顯示，如果應徵者多次的接受問卷調查，回答內容很有可能可以被刻意操弄，甚至因為多次的練習，營造出更為不同的結果，導致結果的與實際狀況不相符；而受試者在人工智慧面試評估系統上，完成前後兩次的非同步視訊面試，所得到的人格特質分析結果，皆沒有產生顯著的變化，顯示出了一致性的得分。因此，當面試官使用人工智慧面試系統分析應徵者的人格特質時，可以獲得更加真實的結果。

關鍵字

非同步視訊面試；五大人格特質；練習效應

並列摘要

Face-to-face employment interviewing is a widely used selection method to identify a job candidate’s personality traits and interpersonal communication skills that are critical criterion for predicting job performance. Other than verbal content, most of interviewers observe a job candidate’s nonverbal responses, and utilized the cues to judge the candidate’s personality traits and communication skills. However, face-to-face interviewing is limited by time and place. Therefore, more and more companies use asynchronous video interviews (AVIs) as an alternative method because AVIs allow the job candidates to answer pre-established questions through a webcam and mobile devices. As advance in artificial intelligence (AI), AVIs enable with AI agent are increasingly adopted as a new automatic interview process by employers. Many job applicants may persist to retake the testing for raise the opportunity to get the job offers. One the one hand the retesting policy merits the procedural justice in personnel selection; on the other hand, it may be a concern for test-retest reliability. Thus, this project aims to examine whether job applicants may change their results through constant practice and retake tests. Applicants may retake the test in order to raise the opportunity to get the job offers. Using new technology and passing appropriate training can enhance the applicant’s control and change of personality traits during the video interview. However, to best of our knowledge, there is no empirical study has been conducted to examine whether such training or practice can allow humans to affect the performance of video interviews by changing their behavior. Accordingly, this study will examined whether the applicant’s assessed personality traits would be changed in test-retest interval change and evaluated the practice effects between traditional self-reported personality traits, other-reported and AI-assessed personality traits. The implications of this study results will be discussed in the conference for improvement and development of industrial education in AI.