Understanding LLM Intervention Explanations in Multi-Party Human-Robot Interaction

2026-06-28

AI Summary

This study addresses the lack of systematic understanding regarding when and why large language model (LLM)-driven robots intervene in multi-human conversations. By comparing dyadic robot–human group interactions under homogeneous versus heterogeneous role configurations—specifically, facilitator versus challenger roles—the authors employ an LLM-based dialogue coordinator, thematic analysis, and controlled experiments to qualitatively code 610 intervention rationales. The work reveals five stable thematic categories that LLMs consistently generate to justify real-time interventions in group dialogue, highlighting consensus-building, inclusive participation, and conversational fluency as core objectives. Furthermore, the findings demonstrate that robot role type significantly shapes intervention strategies: facilitators prioritize process coordination, whereas challengers focus on task progression.

Abstract

Large Language Models (LLMs) are increasingly embedded in social robots to support natural group interactions, yet their role in complex multi-party settings remains underexplored. In particular, it is unclear how LLM-driven robots decide when and why to intervene in group conversations. This paper investigates the intervention explanations generated by an LLM-based orchestrator in a multi-party interaction involving three human participants and two robots. We conducted a between-subjects study with 24 groups (66 university students), comparing a homogeneous condition (two robots with the same role, i.e., a mover) and a heterogeneous condition (two robots with different roles, i.e., a mover and an opposer). At each conversational turn, the LLM orchestrator decided whether to intervene and generated a textual explanation of its decision. We performed a thematic analysis of 610 intervention explanations, identifying five recurring themes. Results show that explanations are facilitation-oriented, emphasizing agreement, participation, and interaction flow. While patterns remain stable across conditions, role differentiation emerges: the mover supports coordination, whereas the opposer drives goal-oriented interventions. These findings contribute to explainable AI by characterizing how LLM-driven systems justify intervention decisions in real-time, multi-party human-robot interaction.

Problem

• Large Language Models
• Human-Robot Interaction
• Multi-Party Interaction
• Intervention Explanations
• Explainable AI

Innovation

• explainable AI
• large language models
• multi-party interaction
• robot intervention
• human-robot collaboration