Towards Designing Better Speech Agent Interaction: Using Eye Gaze for Interaction
Researchers have made significant progress in understanding how to improve interactions with conversational user interfaces (CUIs) by incorporating elements of human-human interaction. A recent study focuses on incorporating abstractions of complex human behavior, specifically gaze, to enhance interactions with speech agents during conversations. The research, conducted across four empirical studies, offers valuable insights for interaction designers and researchers.
One of the key contributions of this research is the development of a novel speech interface called Tama, designed to explore the use of gaze in conversational interactions with smart speakers. The interface, being gaze-aware, aims to leverage the opportunities presented by gaze as a modality to enhance the interaction with the speech agent. However, the studies also highlight the limitations and challenges associated with using gaze as a mode of interaction.
In addition to the development of Tama, the research delves into the empirical contributions documenting interactions with and around speech interfaces, including spontaneous, non-system-directed speech. Through moment-by-moment analysis, the researchers identify the strengths and weaknesses of incorporating gaze as a modality in speech agent interactions.
Furthermore, the research offers a conceptual contribution by providing perspective on minimal anthropomorphic design. By leveraging the skills used in human interactions, designers can advance interactions with speech agents, even if the resulting interactions are not entirely human-like in terms of behavior.
The study emphasizes the need for collaboration between different fields, such as conversation analysis, human-computer interaction, and human-robot interaction, to combine models and approaches from these areas. This collaborative effort would provide essential guidance to designers in building efficient speech systems.
The research identifies three competing yet complementary interaction paradigms across CUIs: Direct Speech Interaction, Agent-Mediated Interaction, and Para-Speech Interaction. Each of these paradigms presents unique challenges and opportunities for interaction designers.
Overall, this research contributes to advancing interactions with speech interfaces and encourages designers to explore the potential of gaze as a modality for enriching conversational interactions. By integrating insights from various disciplines, designers can create more effective and engaging speech systems.
With the possibilities offered by this research, designers and researchers can look forward to designing CUIs that enhance the user experience and provide more seamless interactions. By combining technological advancements with the understanding of human behavior, the next generation of speech interfaces holds promising potential for improving human-technology conversations.
This groundbreaking research demonstrates the importance of incorporating elements of human-human interaction into the design of conversational user interfaces. By leveraging the nuances of gaze and drawing from interdisciplinary approaches, designers can create speech systems that are more intuitive, responsive, and capable of delivering an enhanced user experience.