Toward Metaphor-Fluid Conversation Design for Voice User Interfaces

TL;DR

The paper addresses the mismatch between fixed, human-centered VUI personas and the diverse expectations users bring to different conversational tasks. It proposes Metaphor-Fluid Design, a context-aware approach that dynamically shifts metaphorical framing across use-contexts and includes non-human and fictional metaphors. Through Study 1, it maps metaphor-context alignment and identifies five metaphor roles; through Study 2, it shows that Metaphor-Fluid VUIs enhance perceived enjoyment, adoption intent, and likability compared to a Default, while trust and perceived intelligence remain comparable. The findings challenge the one-size-fits-all paradigm and highlight personalization as a key to effective, adaptive human-AI conversations with practical implications for future VUI and CUI design across modalities.

Abstract

Metaphors play a critical role in shaping user experiences with Voice User Interfaces (VUIs), yet existing designs often rely on static, human-centric metaphors that fail to adapt to diverse contexts and user needs. This paper introduces Metaphor-Fluid Design, a novel approach that dynamically adjusts metaphorical representations based on conversational use-contexts. We compare this approach to a Default VUI, which characterizes the present implementation of commercial VUIs commonly designed around the persona of an assistant, offering a uniform interaction style across contexts. In Study 1 (N=130), metaphors were mapped to four key use-contexts-commands, information seeking, sociality, and error recovery-along the dimensions of formality and hierarchy, revealing distinct preferences for task-specific metaphorical designs. Study 2 (N=91) evaluates a Metaphor-Fluid VUI against a Default VUI, showing that the Metaphor-Fluid VUI enhances perceived intention to adopt, enjoyment, and likability by aligning better with user expectations for different contexts. However, individual differences in metaphor preferences highlight the need for personalization. These findings challenge the one-size-fits-all paradigm of VUI design and demonstrate the potential of Metaphor-Fluid Design to create more adaptive and engaging human-AI interactions.

Figures (4)

• Figure 1: A total of 93 metaphors were identified from Source A (Desai_Twidale_2023), and 225 metaphors were collected from Source B (Chin2024Like). After removing duplicates, 98 unique metaphors remained. From these, 20 metaphors were selected based on their frequency of occurrence and feasibility for practical implementation in VUI design.
• Figure 2: Mapping metaphors and conversational use-contexts on a 7-point scale along the dimensions of formality and hierarchy
• Figure 3: Box plots comparing Metaphor-Fluid VUI and Default VUI across different perception measures: Perceived Enjoyment, Perceived Intention to Adopt, Perceived Trust, Perceived Likability, and Perceived Intelligence. Each box plot displays the distribution of ratings for each measure under both conditions. The blue horizontal line represents the median rating, while the black diamond marker indicates the mean rating. Outliers are represented by individual points outside the whiskers.
• Figure 4: Comparison of Metaphor Use: All vs Unique for Metaphor-Fluid VUI and Default VUI