Collaborative Research: A Process-Driven Approach to Artificial Intelligence Chatbot Interviews

About This Grant

The aim of this project is to study and improve how Artificial Intelligence (AI) chatbots evaluate job candidates. AI chatbots increasingly are used in workplace settings to interview job candidates, offering efficiency and standardization in hiring. AI-based interview systems may unintentionally rely on irrelevant information, however, leading to inappropriate outcomes. This research investigates how AI systems might produce different outcomes based on individual characteristics, even when qualifications are equal. It also explores how people perceive the balance and transparency of such AI interview experiences. The findings inform the development of more robust AI systems and support the deployment of ethical AI in hiring practices, ultimately contributing to a stronger workforce. The project trains students in responsible AI, offers outreach through public forums, and develops interactive dashboards to help human resource professionals make better use of AI tools in hiring. The research in this project analyzes AI-based interview systems through the lens of predictors (e.g., language model embeddings), outcomes (e.g., scores or hiring decisions), and user perceptions (e.g., trust). Drawing on an existing conceptual framework and psychometric natural language processing methods, the research team examines differential functioning of AI predictors across groups, detecting group differences in outcomes, and evaluating candidate reactions to chatbot interviews. Data from both university seniors and working professionals are collected to ensure generalizability. By integrating expertise from psychology, machine learning, and business analytics, the project produces validated metrics, statistical models, and explainable AI tools that enhance transparency and balance in AI-chatbot-based interview systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.