Conversational User Interfaces:

A Workshop on New Theoretical and Methodological Perspectives for Researching Speech-based Conversational Interactions 

Goals

Speech interfaces (aka Voice User Interface [VUIs]) are now commonplace, driven by the wide proliferation of intelligent personal assistants like Amazon Alexa and Google Assistant in mobile and standalone home-based devices. Despite the technical knowledge underpinning speech interface technology advancing rapidly in recent years, research on the user side is limited by comparison. Furthermore, work in the field is highly fragmented, both in the topics studies and the methods used to explore user interactions. Recent research has highlighted significant gaps in theoretically grounded and human-centred speech interface research. Although work as highlighted the types of language choices made by users in speech interface interactions, there is currently no consistent mapping or development of theory to explain these language choices in the literature.

To resolve this we may need to look to other disciplines and existing theories in related fields. Across speech research in HCI, some theoretical concepts and methods have started to be borrowed from other disciplines to explain user behaviour and inform speech interface design (e.g. perspective taking and alignment from psycholinguistic literature). Studies adopting these theoretical approaches tend to borrow methods from the disciplines where these core concepts originiated. Although this allows for the comparison of findings to other research disciplines, the contexts and behaviours of speech interface interaction may not be directly relevant to the way these concepts are traditionally research.

It remains unclear what existing theories can be applied from human-human communication, how they can be tested and explored in an HCI domain and how they may need to be re-conceptualised for this type of interaction. This part-day collaborative workshop focuses on theories and methods we can use to understand user behaviours with speech interfaces. Through the workshop we will discuss:

  • THEORIES - establish what theories, concepts and paradigms are important in understanding our interaction choices with speech interfaces, and potential new theoretical frameworks required to explore them. 

  • METHODS - we will critically evaluate existing methodologies used to explore speech interface interactions and determine priority areas for improving methods to explore theories and paradigms identified.