Human-Like Embodied AI Interviewer: Employing Android ERICA in Real International Conference

TL;DR

This work introduces a human-like embodied AI interviewer that leverages android ERICA and teleoperated TELECO ECAs to conduct qualitative interviews with advanced listening, repair, and user fluency adaptation. A modular dialogue manager coordinates language understanding, backchanneling, and conversation repair, while a post-interview processing pipeline uses chained LLMs to process data and generate presentation materials. Real-world evaluation at SIGDIAL 2024 (n=42) reports 69% positive experiences, demonstrating practical feasibility and marking the first such deployment at an international conference. The study also identifies limitations, including repetition from template-based prompts and reliance on speech-only input, guiding future work toward richer multimodal sensing and culturally adaptive interaction.

Abstract

This paper introduces the human-like embodied AI interviewer which integrates android robots equipped with advanced conversational capabilities, including attentive listening, conversational repairs, and user fluency adaptation. Moreover, it can analyze and present results post-interview. We conducted a real-world case study at SIGDIAL 2024 with 42 participants, of whom 69% reported positive experiences. This study demonstrated the system's effectiveness in conducting interviews just like a human and marked the first employment of such a system at an international conference. The demonstration video is available at https://youtu.be/jCuw9g99KuE.

Figures (15)

• Figure 1: Overall architecture of human-like interview system
• Figure 2: Overall architecture of interview system response generation
• Figure 3: Overall architecture of interview dialogue flow
• Figure 4: Overall architecture of post-interview processing workflow
• Figure 5: Photo of interview dialogue with ERICA by SIGDIAL participant