What Could Possibly Go Wrong When Interacting with Proactive Smart Speakers? A Case Study Using an ESM Application

要旨

Voice user interfaces (VUIs) have made their way into people's daily lives, from voice assistants to smart speakers. Although VUIs typically just react to direct user commands, increasingly, they incorporate elements of proactive behaviors. In particular, proactive smart speakers have the potential for many applications, ranging from healthcare to entertainment; however, their usability in everyday life is subject to interaction errors. To systematically investigate the nature of errors, we designed a voice-based Experience Sampling Method (ESM) application to run on proactive speakers. We captured 1,213 user interactions in a 3-week field deployment in 13 participants' homes. Through auxiliary audio recordings and logs, we identify substantial interaction errors and strategies that users apply to overcome those errors. We further analyze the interaction timings and provide insights into the time cost of errors. We find that, even for answering simple ESMs, interaction errors occur frequently and can hamper the usability of proactive speakers and user experience. Our work also identifies multiple facets of VUIs that can be improved in terms of the timing of speech.

著者
Jing Wei
University of Melbourne, Melbourne, Victoria, Australia
Benjamin Tag
University of Melbourne, Melbourne, Victoria, Australia
Johanne R. Trippas
The University of Melbourne, Melbourne, Victoria, Australia
Tilman Dingler
University of Melbourne, Melbourne, Victoria, Australia
Vassilis Kostakos
University of Melbourne, Melbourne, Victoria, Australia
論文URL

https://dl.acm.org/doi/abs/10.1145/3491102.3517432

動画

会議: CHI 2022

The ACM CHI Conference on Human Factors in Computing Systems (https://chi2022.acm.org/)

セッション: Mouth-based Interaction

290
4 件の発表
2022-05-03 01:15:00
2022-05-03 02:30:00