"You talking to me?" - Exploring voice in self-service user interfaces

Automated teller machines (ATMs) are a classic example of ubiquitous comput ing as they pervade our everyday life and, for many, are typical artifacts of convenient, modern living. More important, most people are unaware that in using an ATM they are connected via a computer to a powerful network. Th rough our research reported within this article, we examine the user accept ance and usability aspects of a novel, "contactless" ATM. This prototype AT M has iris identification, speech recognition, speech synthesis, and commun ication to a user's personal digital assistant (PDA). In this article, we d escribe our experiences of and explorations with "Stella," a contactless AT M. First, we briefly discuss the background to the study in terms of techno logy and trends, and then overview previous usability research in the area. We report our findings from a cognitive walkthrough, the heuristic review of the prototype, focus groups, and a user trial. These results underline t he need for multiple evaluation methods for novel concepts. In general, fro m the findings of this work, consumers believe that speech is a technology that will create more usable applications, even when faced with the relativ e failure of current technology to live up to their expectations or fulfill their practical needs. People are often initially resistant to the concept of iris identification and the PDA interaction, yet basic experience with a usable prototype quickly overcame people's reservations. Finally, we outl ine some of the underlying challenges facing voice-based interaction approa ches to this form of ubiquitous, public self-service computing.