嘉宾介绍：
范津砚，华东师范大学心理学学士（1994年）、硕士（1997年），美国俄亥俄州立大学工业与组织心理学博士（2004年），目前任美国奥本大学心理系教授。主要研究领域是人工智能、人事选拔、新员工入职培训和社会化过程、和跨文化适应和培训。他在组织行为与组织心理学的专业杂志上（比如Journal of Applied Psychology, Journal of Organizational Behavior, Journal of Management）发表了多篇论文，获得了美国工业与组织心理学会（厂滨翱笔）和美国管理学会（础辞惭）的一系列奖项和基金资助，曾担任Journal of Vocational Behavior杂志的副主编（2019 – 2021），目前是 Journal of Applied Psychology和Journal of Organizational Behavior的编委会成员。在实践方面，范博士开发了一系列人才测评的工具、模型、方法，并长期从事人力资源管理相关的公司咨询工作。
讲座介绍：
Personality inventories are commonly used in talent assessments, but traditional self-reports are susceptible to faking, especially in high-stakes situations. This study compared the faking resistance of machine-inferred personality scores derived from an AI-chatbot with self-reported and interviewer-rated personality scores. Participants were 398 U.S. college students. Half of them were instructed to fake responses on a traditional self-report personality measure and a text-based interview through an AI chatbot for a desirable internship, while another half were instructed to respond honestly. Machine-inferred scores were calculated using pre-trained deep-learning and bag-of-words models. Trained interviewers read participants’ chat scripts and rated their personality. Results indicated that machine-inferred personality scores were the most faking resistant, followed by interviewer rated personality scores, then self-reported personality scores. Limitations, future research directions, and practical implications will be discussed.
