← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence medium-high

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

Zhihan Jiang , Qianhui Chen , Chu Zhang , Yanheng Li , RAY LC

This is a thoughtful CHI paper that reframes silence from a bug into an interaction resource. The contribution is strongest as an interaction-design framework backed by a controlled study, with clear evidence that context-aware pacing can improve perceived listening and engagement in supportive text conversations.


Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Contribution shape

Knowledge form
generative knowledge typical · 35/268
Novelty type
framework typical · 59/268
Abstraction level
interaction typical · 22/268
Generalization target
design family typical · 38/268
Validation mode
controlled experiment typical · 47/268

Evidence profile

Evidence strength
moderate typical · 105/268
Claim alignment
strong typical · 231/268
Overclaim risk
medium typical · 210/268

Review Summary

This paper’s main value is conceptual: it challenges the common-sense assumption that conversational agents should optimize for immediacy and verbal throughput, and instead argues that timing itself can communicate attention, empathy, and listening. That is a meaningful shift for CHI because it moves the design lens from content generation to temporal interaction cues. The paper’s novelty is not just that it uses pauses, but that it systematizes them into five context-aware pacing strategies and then evaluates the idea in a controlled between-subjects study. The evidence quoted in the paper supports a bounded but credible claim: in two supportive scenarios, the context-aware agent performed better than a static-pacing control on human-likeness, smoothness, interactivity, and some downstream relational outcomes such as listening quality and affective trust. At the same time, the paper is careful enough to acknowledge important limits: the study is text-only, silence can be ambiguous without multimodal cues, and the scenarios are predefined role-play settings rather than open-ended deployment. So I would read the contribution as a strong interaction-design and empirical finding package, not as a universal prescription for all conversational agents. The work is especially compelling for reflective, emotionally supportive contexts where a pause can plausibly be interpreted as active listening rather than latency. Its broader field significance is that it legitimizes pacing as a first-class design variable in human-AI communication, which is a useful and publishable CHI direction.

What Changed

Canon before

Conversational agents are typically designed around prompt, efficient turn-taking and verbal output; silence is often treated as delay or failure rather than an interaction resource.

Departure from common sense

The paper argues against the default assumption that supportive conversational agents should respond as quickly as possible. Instead, it treats silence and pacing as meaningful signals for active listening, suggesting that deliberate pauses can improve perceived empathy and engagement rather than merely harming responsiveness.

Actual novelty

The paper claims a systematic contribution to conversational agent design by introducing conversational pacing as a design concern and operationalizing active listening through five context-aware pacing strategies. It also frames the work as providing empirical evidence that strategic silence can support empathic human-AI relationships.

Evidence

The paper combines qualitative analysis of ten active-listening cases with a between-subjects study of 50 participants across two supportive scenarios. The reported results indicate that the context-aware pacing agent outperformed a static-pacing control on human-likeness, smoothness, interactivity, and some trust/listening measures, but the evidence is bounded to text-based, scenario-based interaction.

“ To the best of our knowledge, we are the first to systematically introduce conversational pacing into CA design, providing empirical evidence that the strategic use of silence for context-aware pacing is critical for developing empathic human-AI relationships analogous to active listening”

actual novelty · Introduction contributions bullet · confidence 0.74

“Index Terms Hear You in Silence: Designing for Active Listening in Human Interaction with Conversational Agents Using Context-Aware Pacing Human-centered computing Collaborative and social computing Empirical studies in collaborative and social computing”

departure from common sense · Abstract/Introduction framing of the problem and proposed approach · confidence 0.78

“ Without multimodal cues, text-based silence can be ambiguous and misinterpreted as system lag rather than thoughtful deliberatio”

limitation · Limitations and Future Work (6.6) · confidence 0.77

“ 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”

validation scope · Abstract and results sections (RQ1/RQ2) · confidence 0.80

Limits

Method limits

The evaluation is limited to a between-subjects study with N=50 and two predefined supportive scenarios, so the evidence supports the proposed interaction design within that setting rather than a broad causal claim about all conversational agents or all forms of empathy.

Deployment limits

The approach is text-based and depends on pacing/silence being legible to users; deployment in richer multimodal or high-stakes settings may require additional cues, stronger control over generation, and careful handling of latency versus intentional silence.

Boundary conditions

The findings are most applicable to supportive, reflective conversations where pauses can be interpreted as attentive listening. They are less certain for fast transactional tasks, multimodal interfaces, or contexts where silence may be mistaken for system failure.

Position in field

This work extends CHI conversations about empathic conversational agents by shifting attention from what agents say to when they say it. Its contribution is best read as an interaction-design framework and empirical demonstration for pacing as a listening cue, rather than a general-purpose dialogue policy.

Abstract