[Paper Review] Hear You in Silence: Designing for Active Listening in Human Interaction with Conversational Agents Using Context-Aware Pacing

Zhihan Jiang, Qianhui Chen+3 more|arXiv (Cornell University)|Feb 5, 2026

Social Robot Interaction and HRI0 citations

TL;DR

The paper designs and evaluates a context-aware pacing framework for conversational agents to implement active listening through strategically timed silence, showing improved perceived interaction quality over static pacing in two supportive scenarios.

ABSTRACT

In human conversation, empathic dialogue requires nuanced temporal cues indicating whether the conversational partner is paying attention. This type of "active listening" is overlooked in the design of Conversational Agents (CAs), which use the same pacing for one conversation. To model the temporal cues in human conversation, we need CAs that dynamically adjust response pacing according to user input. We qualitatively analyzed ten cases of active listening to distill five context-aware pacing strategies: Reflective Silence, Facilitative Silence, Empathic Silence, Holding Space, and Immediate Response. In a between-subjects study (N=50) with two conversational scenarios (relationship and career-support), the context-aware agent scored higher than static-pacing control on perceived human-likeness, smoothness, and interactivity, supporting deeper self-disclosure and higher engagement. In the career support scenario, the CA yielded higher perceived listening quality and affective trust. This work shows how insights from human conversation like context-aware pacing can empower the design of more empathic human-AI communication.

Motivation & Objective

• Systematically study how pacing (timing of responses) functions as an active listening cue in human-AI conversations.
• Identify context-aware pacing strategies that reflect and respond to user emotions and information needs.
• Translate insights from human conversations into implementable CA designs.
• Empirically evaluate the impact of context-aware pacing on user experience and disclosure in two scenarios.

Proposed method

• Qualitative analysis of ten real-world active listening cases to identify pacing strategies.
• Development of an LLM-based conversational agent that implements context-aware pacing with five strategies: Reflective Silence, Facilitative Silence, Empathic Silence, Holding Space, and Immediate Response.
• Operationalization of strategies into eight concrete tactics (e.g., Recognize, Reconfirm, Reengage, Reposition, Reconsider, Resonate, Holding, Resolve).
• Between-subjects experimental study (N=50) comparing context-aware pacing CA against a static-pacing CA across two supportive scenarios (career and relationship).
• Quantitative assessment of perceived interaction quality (listening quality, affective/cognitive trust, human-likeness, smoothness, interactivity) and interaction behaviors (depth of self-disclosure, level of engagement).
• Analysis of pacing distribution and transition patterns across cases to derive design guidance.

Experimental results

Research questions

• RQ1 RQ1: How do context-aware pacing strategies in CAs affect users’ perceived interaction quality and experience (listening quality, affective trust, cognitive trust, human-likeness, smoothness, interactivity) and engagement?
• RQ2 RQ2: How does context-aware pacing influence user interaction behaviors in text-based supportive conversations with CAs (depth of self-disclosure and level of engagement)?

Key findings

• Context-aware pacing CA scored higher than a static-pacing control on perceived human-likeness, smoothness, and interactivity.
• In the career-support scenario, the context-aware CA yielded higher perceived listening quality and affective trust.
• Five context-aware pacing strategies were identified and operationalized for implementation: Reflective Silence, Facilitative Silence, Empathic Silence, Holding Space, and Immediate Response.
• Pacing was distributed with Resolve and Reconfirm appearing most frequently; Holding and other high-intensity strategies were used selectively depending on the emotional intensity of the input.
• Two pacing transition patterns emerged: Conversational Inertia (information-focused moves persist) and Supportive Arcs (emotional regulation sequences).