The increasing volume of job applications received by organizations presents significant logistical and ethical challenges for recruitment processes. Traditional methods of screening and interviewing are often inefficient, inconsistent, and susceptible to bias, especially at scale. In this work, we introduce the Multi-Agent Interviewing System (MAIS), an AI-powered interview platform designed to support scalable and fair recruitment while preserving the human experience at the center of the process. MAIS combines multilingual, multi-channel communication with real-time monitoring and safeguards against misuse, enabling candidates to interact naturally with an AI interviewer through speech or text. We conducted a two-part user study involving 22 participants to explore both the usability of MAIS and the human perceptions that arise during interaction with an AI interviewer. Participants reported high usability and engagement, while qualitative feedback revealed themes of trust, transparency, and perceived fairness. These findings highlight not only the potential of MAIS to reduce recruiter workload and enhance consistency but also provide new insights into human-AI interaction dynamics in evaluative, high-stakes contexts. This work contributes to the design of safer, more transparent, and human-centered generative AI applications in recruitment and beyond.
Piras et al. (Mon,) studied this question.