I spent last summer writing my Master’s dissertation on practitioners’ approach to design guidelines for voice user interfaces (VUIs) such as Alexa, Google Assistant, etc. During that time, I realized there is limited research based on Conversational AI practitioner communities and the work they do. In particular, the design and development practices they follow while building VUIs.
Horizon Digital Economy Research and the Mixed Reality Lab at the University of Nottingham have been researching voice interfaces for the last few years – see work by Stuart Reeves, and Joel Fischer and Martin Porcheron – now at Swansea University. After a discussion with Stuart about potential areas of research in this domain, I was offered an internship to carry out a study to explore the collaboration between designers and developers who build conversational user interfaces (CUIs).
The research I conducted for my MSc dissertation provided foundational knowledge of the design practices practitioners employ while designing for VUIs. I learned more about the development practices by engaging with practitioner-based communities such as Voice Lunch. Designated members of Voice Lunch host a Zoom session on a weekly basis. This allows them to connect with other practitioners in the space and stay up to date with advancements within the field. In the first half of the session, a member of the practitioner community presents something relevant to the theme of the session (i.e., Development, Language & Linguistics, etc.). The remainder of the session is used to discuss the presentation, addressing questions to the presenter, and sharing thoughts.
Attending these Voice Lunch sessions – especially the ones focused on development – helped gain insight into what development for VUIs might look like, and challenges practitioner communities face, for example, lack of tools, limited support from platforms. It also provided more exposure to experts working in this domain and facilitated building connections, which helped with participant recruitment for my internship study.
During my 10-month internship, I co-developed my research project alongside my supervisor – Stuart Reeves. While I worked from home, we introduced regular virtual meetings, during which I provided updates on the progress of the research, discussed concerns and talked through ideas based on the findings of our interview-based study. Furthermore, I worked towards publishing papers for conferences, submitting one based on my Master’s dissertation to the ACM SIGCHI conference. I’m currently working on a paper based on my internship.
The Horizon internship provided insight into what conducting research for a PhD involves. Furthermore, being proactive about using different methods to gain a deeper understanding of the research topic has fuelled my curiosity about Conversational AI. Additionally, the opportunity to co-develop my research and drive it forward has helped me improve my research skills, specifically being able to review literature, conduct interviews and analyse data qualitatively.
Written by Krishika H. Khemani
Tags: AI, Alexa, Artificial Intelligence, Google Assistant, voice user interfaces