Voice User Interfaces

The commercial availability and acceptability of voice user interfaces (VUIs) has increased greatly in recent years and the research looking at VUIs has increased in tandem. As a sociophonetician (someone who studies the social factors that influence the production and perception of speech) I am interested in the voices that these devices use, how they make users react and feel, and the process by which they are designed and selected. However, from an initial survey of the literature it appears that these topics have not been approached yet. This made me consider all the sociophonetic knowledge and theory that could be useful in this context, and I was inspired to explore this further.

The initial result of considering this topic is the paper “Voice as a design material: Sociophonetic Inspired Design Strategies in Human-Computer Interaction“; a conceptual essay where I and my co-authors recommend design approaches for the design of VUI voice output by introducing sociophonetics into HCI. Thus, we define a new research area and set it’s agenda. The paper is written for a HCI audience, but for linguists that might be interested I’ve also written a very brief introduction to HCI below to provide some background and context.

Future Work

This paper is just the start of this research topic and it opens up a variety of directions that could be taken. First, I would like to explore the design approaches that we recommend in the paper and consider the methods by which they can be implemented. This will probably involve looking at how current design resources and tool kits can be translated from other mediums into voice, how non-linguists can talk about and describe voice, and I’d like to explore how design fictions can be used as well. There are also many other research topics that come to light when you overlap the interests of sociophonetics and HCI in the context of VUIs. Other questions I’d like to ask include ‘do users accommodate to the voice of the VUI?’ and ‘is user perception of the VUI voice influenced by the physical design of the device?’

A very brief introduction to HCI

My personal, somewhat longwinded, definition of human-computer interaction (HCI) is ‘a research field is that examines how the design and use of technology impacts on the individual, social groups, and society in the present and the future from physical, social, cultural, ethical, and economical perspectives.’

HCI is difficult to define because it is forever changing at rapid speed. Every time a new technology is developed, is placed into a new context, or viewed from a new perspective, there is the opportunity for another outside discipline or research field to be integrated. This brings new methods, theories and epistemological perspectives. As a result, the boundaries of what is and is not HCI are continually expanding. Therefore, at CHI (ACM Conference on Human Factors in Computing Systems 2019) the venue where this piece is published, the types of papers range from examining ergonomics through experiments, to philosophical and ethical musings. From my experience so far, it seems HCI isn’t something a researcher is, unlike in many other disciplines (e.g. “I’m a linguist / historian / biologist”) but something you do (e.g. “I’m a designer / psychologist / computer scientist that researches X”, “X” being a specific topic being examined through a HCI lens).

The kind of HCI research that this paper considers is design methods – how choices are made, in practical terms and from a theoretical perspective, when a piece of technology is being developed. The premise is that improving design methods will allow for more appropriate technology to be developed and lead to better user experiences.